NETWORK DYNAMICS OF THE AMERICAN FAR-RIGHT CRIMINAL ANTI-TAX MOVEMENT By Brandon A. Sullivan A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Criminal Justice - Doctor of Philosophy 2015 ABSTRACT NETWORK DYNAMICS OF THE AMERICAN FAR-RIGHT CRIMINAL ANTI-TAX MOVEMENT By Brandon A. Sullivan This study examines the network dynamics of far-right extremists linked to the anti-tax movement in the United States engaging in financial crime schemes from 1990 to 2010. Data for the study are drawn from the United States Extremist Financial Crime Database (EFCDB), an open source relational database focusing on financial schemes by political and religious extremists, including far-right extremist tax protesters. Mixed method, multilevel, longitudinal network analysis is used to examine the structure and characteristics of individual criminal perpetrators participating in financial schemes. After examining both lone offender and multisuspect schemes and different motivations for engaging in financial crimes, cohesive subgroups in a two-mode network are identified and analyzed over time. Subsequent qualitative analysis of connections between different cohesive subgroups revealed important themes underlying the changing composition of the network centered around four large anti-tax organizations. This study enhances the current understanding of dynamic motivations, behaviors, and network structures of financial crime schemes involving anti-tax extremist criminals and their nonextremist collaborators. Copyright by BRANDON A. SULLIVAN 2015 Dedicated to the memory of Ron “Stone Cold” Sullivan. iv ACKNOWLEDGEMENTS I would like to acknowledge the support of the National Consortium for the Study of Terrorism and Responses to Terrorism (START). The funding and resources provided by the START Terrorism Research Award (TRA) in addition to ongoing support of the U.S. Extremist Crime Database (ECDB) study were an invaluable contribution to the completion of this dissertation. Numerous individuals contributed to the successful completion of this dissertation. First, a special thank you to Dr. Steve Chermak whose consistent support, patience, and guidance made the completion of this dissertation and doctorate degree possible. I will always be grateful for everything you have done. I would also like to thank the other members of my committee who were essential not only to improving this dissertation, but also to my professional and intellectual development. Dr. Joshua Freilich, for constant consultation and insight in the development of the Financial Crimes section of the ECDB and the numerous exciting studies being developed from these efforts. I appreciate both Dr. Chermak and Dr. Freilich for trusting me with the responsibility for taking on this enormous endeavor and for allowing me to use these data for this dissertation. I hope the outcome has exceeded expectations and I look forward to many years of future collaboration. Dr. Jeremy Wilson and Dr. Ed McGarrell, whose advice and encouragement have been crucial to enhancing my research efforts throughout my career. Without your support, this entire endeavor would not have been possible. Finally, Dr. Ken Frank, whose workshops and conversations formed the intellectual foundation for the network analyses in this dissertation. Thank you for providing invaluable guidance in working through the various aspects of this research. It will undoubtedly serve me well far into the future. v I would like to offer a special thank you to members of the ECDB team who worked hard over the last five years coding and searching cases to prepare the Financial Crime data used in this dissertation, especially Dr. Ashmini Kerodal, Colleen Mills, Dr. Brittany Hayes, Dr. Katharine Boyd, Joel Capellan, Angela Juliano, and Sarah Fitzgerald. Also thank you to Dr. Roberta Belli, Dr. Jeff Gruenewald, and Dr. William Parkin for their tremendous effort developing and maintaining the ECDB. To my friends and colleagues in the School of Criminal Justice, too numerous to name here, who played a part in making this journey fulfilling and worthwhile, thank you for all the memories these last few years. Finally, to my family, who sacrificed the most to see this journey through to conclusion. I appreciate you putting up with all those times I was not there, either physically or mentally. Most of all, my wife Katie, whose incredible love, tolerance, and encouragement never wavered despite the seemingly constant barrage of challenges. I look forward to beginning the next stage of the journey together. vi TABLE OF CONTENTS LIST OF TABLES .......................................................................................................................... x   LIST OF FIGURES ....................................................................................................................... xi   CHAPTER 1: INTRODUCTION ................................................................................................... 1   Research Relevance ............................................................................................................ 3   Far-Right Terrorism and Extremism ....................................................................... 3   Organizational Aspects of Financial Crime ............................................................ 5   Far-Right Extremists and Financial Crime ............................................................. 7   Network Analysis and Extremist Financial Crime ................................................. 9   Policy Implications ............................................................................................... 10   Research Project................................................................................................................ 11   Research Questions ........................................................................................................... 12   Data, Methodology, and Analysis ..................................................................................... 13   Overview of Subsequent Chapters .................................................................................... 16   CHAPTER 2: LITERATURE REVIEW ...................................................................................... 17   White-Collar and Organized Crime .................................................................................. 17   Far-Right Extremist Financial Crime................................................................................ 20   American Anti-Tax Movement ......................................................................................... 22   Frivolous Tax Arguments ..................................................................................... 25   Connections to Sovereign Citizen Movement ...................................................... 29   Structure of Criminal Activities by Far-Right Extremists ................................................ 31   Lone Wolves ......................................................................................................... 32   Co-Offending Networks........................................................................................ 34   Conclusion ........................................................................................................................ 37   CHAPTER 3: THEORY ............................................................................................................... 39   Development of Extremist Beliefs .................................................................................... 39   From Belief to Behavior ................................................................................................... 43   Conclusion ........................................................................................................................ 51   CHAPTER 4: METHODS ............................................................................................................ 52   Research Design................................................................................................................ 52   Data ....................................................................................................................... 52   Sample Selection................................................................................................... 54   Data Collection ................................................................................................................. 58   Open Source Searching ......................................................................................... 58   Coding ................................................................................................................... 60   Financial Scheme ...................................................................................... 61   Criminal Suspect ....................................................................................... 62 Ideological and Greed Association ....................................................................... 63   Reliability.............................................................................................................. 66   Analysis............................................................................................................................. 68   vii Descriptive Analysis ............................................................................................. 68   Two-Mode Social Network Analysis.................................................................... 70   Cohesive Subgroups.............................................................................................. 73   Network Visualization .......................................................................................... 75   Qualitative Analysis .............................................................................................. 76   Conclusion ........................................................................................................................ 78   CHAPTER 5: DESCRIPTIVE ANALYSIS ................................................................................. 79   Financial Schemes ............................................................................................................ 79   Size........................................................................................................................ 80   Length ................................................................................................................... 82   Type ...................................................................................................................... 83   Motive ................................................................................................................... 86   Criminal Suspects ............................................................................................................. 90   Demographics ....................................................................................................... 91   Affiliation.............................................................................................................. 93   Motive ................................................................................................................... 95   Conclusion ...................................................................................................................... 101   CHAPTER 6: COHESIVE SUBGROUP ANALYSIS .............................................................. 102   Overview of Data Structure ............................................................................................ 102   Results of Cohesive Subgroup Analysis ......................................................................... 110   Cohesive Subgroup Descriptions .................................................................................... 111   Longitudinal Analysis of Cohesive Subgroups............................................................... 117   Implications..................................................................................................................... 119   Directions for Additional Research ................................................................................ 121   Conclusion ...................................................................................................................... 122   CHAPTER 7: QUALITATIVE ANALYSIS ............................................................................. 124   Evaluation of Original Source Materials ........................................................................ 124   Expanding Network Beyond Cohesive Subgroups ......................................................... 129   Evolution of Network ..................................................................................................... 135   Implications..................................................................................................................... 138   Limitations ...................................................................................................................... 140   Conclusion ...................................................................................................................... 140   CHAPTER 8: DISCUSSION...................................................................................................... 142   Overview of Results........................................................................................................ 142   Theoretical Implications ................................................................................................. 145   Methodological Implications .......................................................................................... 151   Policy Implications ......................................................................................................... 154   Conclusion ...................................................................................................................... 158   APPENDICES ............................................................................................................................ 159   Appendix A. Sources Reviewed by the ECDB to Identify Illegal Violent Incidents and Financial Schemes .......................................................................................................... 160   Appendix B. Open Source Searching Protocol ............................................................... 163   viii Appendix C. Ranking of Source Reliability ................................................................... 165   Appendix D. Strength of Ideological Association .......................................................... 166   Appendix E. Strength of Greed Motive .......................................................................... 167   Appendix F. Figure A.1. Network Evolution Sociograms of Cohesive Subgroups Illustrating Ideological Motivation in Far-Right Criminal Anti-Tax Movement ........... 168   Appendix G. Figure A.2. Network Evolution Sociograms of Cohesive Subgroups Illustrating Greed Motivation in Far-Right Criminal Anti-Tax Movement.................... 179   Appendix H. Figure A.3. Network Evolution Sociograms with Anti-Tax Organizations Illustrating Ideological Motivation ................................................................................. 190   Appendix I. Figure A.4. Network Evolution Sociograms with Anti-Tax Organizations Illustrating Greed Motivation ......................................................................................... 206   REFERENCES ........................................................................................................................... 222   ix LIST OF TABLES Table 4.1. Association between Membership in the Same Cluster and Participation in a Financial Scheme ........................................................................................................................73 Table 5.1. Scheme Characteristics ..............................................................................................79 Table 5.2. Number of Suspects Involved in Financial Schemes.................................................81 Table 5.3. Financial Scheme Type: Lone Wolf vs. Multi-Suspect .............................................84 Table 5.4. Financial Scheme Motive: Lone Wolf vs. Multi-Suspect..........................................86 Table 5.5. Scheme Strength of Ideology and Greed Association ...............................................87 Table 5.6. Scheme Strength of Ideology Association: Lone-Wolf vs. Multi-Suspect ................88 Table 5.7. Scheme Strength of Greed Association: Lone-Wolf vs. Multi-Suspect ....................89 Table 5.8. Suspect Characteristics ..............................................................................................90 Table 5.9. Suspect Motives .........................................................................................................96 Table 5.10. Suspect Strength of Ideology and Greed Association .............................................97 Table 5.11. Suspect Strength of Ideology Association: Far-Right vs. Non-Extremist ...............98 Table 5.12. Suspect Strength of Greed Association: Far-Right vs. Non-Extremist .................100 Table 6.1. Network Components ..............................................................................................102 Table 7.1. Scheme Characteristics for Updated Network .........................................................126 Table 7.2. Suspect Characteristics for Updated Network .........................................................127 x LIST OF FIGURES Figure 4.1. Financial Schemes Involving Tax Protesters by Year: 1990 to 2010 ......................58 Figure 4.2. Equation for Identifying Cluster Membership in Two-Mode Network ...................74 Figure 5.1. Active Schemes by Year: 1990 to 2010 ...................................................................83 Figure 5.2. State of Suspect Primary Residence .........................................................................92 Figure 6.1. Whole Network Illustrating Ideological Motivation: All Components..................104 Figure 6.2. Whole Network Illustrating Greed Motivation: All Components ..........................105 Figure 6.3. Whole Network Illustrating Ideological Motivation: Two and Three-Node Components Removed ..............................................................................................................107 Figure 6.4. Whole Network Illustrating Greed Motivation: Two and Three-Node Components Removed ...................................................................................................................................108 Figure 6.5. Network Sociograms of Cohesive Subgroups Illustrating Ideological Motivation in Far-Right Criminal Anti-Tax Movement: 1990-2010 ..............................................................112 Figure 6.6. Network Sociograms of Cohesive Subgroups Illustrating Greed Motivation in FarRight Criminal Anti-Tax Movement: 1990-2010 .....................................................................113 Figure 7.1. Initial Network Sociogram Illustrating Ideological Motivation .............................130 Figure 7.2. Initial Network Sociogram Illustrating Greed Motivation .....................................131 Figure 7.3. Network Sociograms with Anti-Tax Organizations Illustrating Ideological Motivation: 1993 to 2008..........................................................................................................133 Figure 7.4. Network Sociograms with Anti-Tax Organizations Illustrating Greed Motivation: 1993 to 2008 .............................................................................................................................134 Figure A.1. Network Evolution Sociograms of Cohesive Subgroups Illustrating Ideological Motivation in Far-Right Criminal Anti-Tax Movement ...........................................................168   Figure A.2. Network Evolution Sociograms of Cohesive Subgroups Illustrating Greed Motivation in Far-Right Criminal Anti-Tax Movement ...........................................................179   Figure A.3. Network Evolution Sociograms with Anti-Tax Organizations Illustrating Ideological Motivation ..............................................................................................................190   xi Figure A.4. Network Evolution Sociograms with Anti-Tax Organizations Illustrating Greed Motivation .................................................................................................................................206   xii CHAPTER 1: INTRODUCTION Financial crimes have received minimal attention compared to other types of crime (e.g. violent incidents). Tax evasion and avoidance in particular remains a persistent problem, but has received little attention in criminology and criminal justice research (Levi, 2010). While the emphasis of tax fraud typically centers on wealthy individuals and large corporations due to the vast scale and fiscal implications of the problem (U.S. GAO, 2008), little attention has been paid to financial crimes involving ideologically motivated tax protesters. It is important to consider how individuals engage in financial crime behaviors together over time, including compliance with tax and other laws (Andrei, Comer, & Koehler, 2014; Kirchler, 2007). Systematic empirical research into connections among tax avoidance, anti-tax, and anti-government beliefs and tax fraud behaviors is necessary for the advancement of scholarship in this area and the development of sound public policies (Leighton, 2010). Analyzing known cases of financial crime involving tax protesters can determine both causes and consequences of these frauds and identify areas of intervention (V. Braithwaite, 2010; Levi, 2010). This study applies network analysis to an area of research receiving limited attention: the far-right extremist criminal anti-tax movement. This study does not focus on tax evaders generally, but instead concentrates on offenders ideologically opposed to taxation who use various frivolous legal arguments to justify their beliefs and behaviors. As opposed to the typical tax evader, the anti-tax movement studied here is conceptualized as a loosely associated array of individuals who believe tax laws are illegitimate or do not apply to them. They are also referred to as tax protesters or tax deniers. They do not simply desire to avoid taxes due to personal greed. While some tax protesters offend for personal financial gain, they typically have more complex motivations. Anti-tax extremists exhibit an ideological grievance against the government that 1 drives their tax refusal. These ideologically motivated offenders believe they have a moral or legal duty to refuse taxation. However, there is a wide array of motivations for engaging in financial crime schemes. Some are indeed ideologically motivated, but also take advantage of widespread distrust of the government, dislike for paying taxes, and ignorance of the law to defraud victims to commit various types of profit motivated financial schemes. Others adhere to different far-right extremist belief systems in addition to being tax protesters, such as those held by sovereign citizens, militia, “Christian Patriots,” white supremacists, neo-Nazis, and Christian Identity proponents. This suggests that these criminals can hold far-right right extremists and carry out schemes with either ideological or greed motivations. In addition, those without extremist affiliations often collaborate with tax protesters to carry out different types of schemes. The complexity of financial crime activities linked to the general anti-tax movement warrants further attention empirically and theoretically and is the main focus of this study. The premise of this study is that while anti-tax crimes are committed by both lone offenders and through links to larger organizations, numerous informal associations of offenders exist within this relatively decentralized anti-tax movement who actively engage in financial crime schemes together. While there is relatively general agreement among scholars and practitioners that the broader far-right extremist movement is growing, systematic data and analysis is scarce. Empirical research into the dynamic motivations, behaviors, and structures of the anti-tax movement over time will develop a richer understanding of these ideologically motivated financial crime networks. This chapter introduces the main conceptual ideas and research strategy for this study. The relevance of the research is outlined first, setting the stage for the research questions. The methodology, including data collection and analytical strategy, are then discussed. The chapter concludes with an overview of the remaining chapters. 2 Research Relevance The first section of this chapter outlines the relevance of the current study. First, the current knowledge on terrorism/extremism and organizational aspects of financial crime are discussed. Second, the limited attention to far-right extremist financial crime is highlighted. Third, the importance of network analysis and co-offending networks is outlined. Lastly, potential policy implications stemming from the current study are introduced. Far-Right Terrorism and Extremism Far-right extremism1 has received little empirical attention in criminal justice and criminology despite the substantial threat to public safety and stability (Chermak, Freilich, & Shemtob, 2009; Freilich & Chermak, 2009; Freilich, Chermak, & Simone, 2009). Scholarly attention to terrorism has increased dramatically since the September 2001 terrorist attacks, evidenced by the influx of research funding, degree programs, journals, and publications related to the study of terrorism. These efforts have resulted in a better understanding of many terrorismrelated issues, but many questions remain unexplored and/or unanswered. This study focuses on one particularly underdeveloped area: far-right extremist financial crime. In addition to financial criminal activities, political and religious extremists continue to promote violence in the United States, causing fatalities and injuries, property damage, and financial losses (Chermak, Freilich, & Simone, 2010). Some commit substantially harmful violent acts, such as deadly attacks on law enforcement officers and the destruction of government buildings (ADL, 2005; Corcoran, 1990; Flynn & Gerhardt, 1995; Freilich & 1 Far-right extremists subscribe to aspects of the following beliefs. They are fiercely nationalistic, anti-global, suspicious of federal authority and reverent of individual liberties, especially their right to own guns and be free of taxes. They believe in conspiracy theories involving imminent threats to national sovereignty or personal liberty and believe that their personal or national ‘way of life’ is under attack. Sometimes such beliefs are vague, but for some the threat originates from specific racial or religious groups. They believe that they must be prepared to defend against this attack by participating in paramilitary training or survivalism (Freilich, Chermak, Belli, Gruenewald, & Parkin, 2014). 3 Chermak, 2009; Hamm, 2005; Hamm & van de Voorde, 2005). Many far-right extremists escalate from committing non-violent offenses, such as tax avoidance, to engaging in violence against citizens, police officers, and other government officials (Freilich & Chermak, 2009). State police report far-right domestic extremist groups to be a persistent threat to public safety (Carter, Chermak, Carter, & Drew, 2014; Chermak et al., 2010). In particular, these violent acts have grown among sovereign citizens, who share many of the same characteristics as tax protesters, including anti-government attitudes and beliefs about the illegality of the tax system. Far-right extremists not only commit a wide array of violent crimes but are also involved in non-violent crime with the intent to cause financial harm. A common harassment and retaliation tactic is to intimidate officials and citizens with false liens and other legal documents, known as “paper terrorism” (ADL, 2012; Barkun, 1997; Flynn & Gerhardt, 1995; Pitcavage, 1998, 1999). The commonly held belief that government officials do not have legal authority over them results in conflicts with police and other government authorities. These frivolous documents are filed against those who they believe have wronged them or other movement members. In addition, far-right extremists are involved in a number of other ideologically motivated financial schemes, including tax avoidance and bank/check fraud. Although domestic terrorism has become a top concern among law enforcement officials, criminology and criminal justice scholars have virtually ignored (with a few exceptions) financial aspects of terrorism. Practitioners and scholars from other disciplines have paid more attention to terrorism financing, but most existing literature is journalistic, focuses on individual case studies, or lacks empirical rigor. The majority of research on extremism and terrorism focuses on a small number of high profile violent incidents while failing to mention financial crimes, material support, or preparatory crimes committed by ideologically motivated offenders 4 (Gruenewald, Freilich, & Chermak, 2009). This is an important omission in the literature. This study confirms that far-right extremists are extensively involved in financial crimes, suggesting that this overlap between terrorism and extremism and financial crime deserves further attention. The next section discusses the overlap between organized criminal activities and financial crime, followed by a review of far-right extremist financial crime. Organizational Aspects of Financial Crime Far-right extremist financial crime is incredibly unique and difficult to situate in the existing literature. It is one of many types of crime that are actually hybrids between what has traditionally been considered organized crime and white-collar crime. These crimes involve legitimate and illegitimate elements, individuals and businesses, working professionals and crime groups. The lines between white-collar crime and organizational crime are often blurred, making distinctions between the two largely irrelevant (Albanese, 2000; Friedrichs, 2004). In many ways the two are intricately related, with criminals engaging in financial crimes together in some organized fashion based on shared opportunity. Extremist financial crime in particular does not fit neatly into either of these categories as a hybrid of organized, white collar, and ideologically motivated crime. It is unique in that the victim is often the government (and more broadly, taxpayers), particularly for cases of tax avoidance. This study lies at the intersection between these two general areas of research by focusing on the organizational aspects of far-right extremist financial crime. The organization of these crimes ranges from larger, more integrated groups to loosely connected networks of individuals with different levels of involvement in crime. Others involve lone wolf offenders who engage in tax avoidance or other schemes that seemingly disconnected from other financial crime activity in the anti-tax movement. These offenders often receive 5 information on how to carry out a financial scheme informally through their network connections, such as with friends, family member, or business associates. Other times individuals obtain anti-tax information over the Internet, using it as the basis to create their own financial schemes. The connections to the broader anti-tax movement influence the ability to successfully engage in financial crimes by facilitating both anti-tax belief and behavior. Some who are deeply interconnected to a few key individuals will form more cohesive subgroups, often centered on additional connections beyond co-offending ties, including family and friend relationships. These group dynamics then have additional influences the beliefs and behaviors of its members. Focusing on these relational and situational elements involved in carrying out different types of financial crime schemes captures the various aspects of how these crimes are organized and who engages in them. Opportunity and motivation are is two of the main driving forces behind the organization of these financial crimes. A person must be in a position to commit a crime by deception, misrepresentation, or manipulation in order for a financial crime to occur. In the case of individual tax avoidance, every U.S. taxpayer has the opportunity to stop filing his or her tax returns or paying taxes. The question is what drives the motivation to engage in this criminal behavior. In this study, the dichotomy between ideology and greed motivations is examined to determine what drives anti-tax behavior. Some individuals may develop anti-tax philosophy first and then decide to stop paying taxes based on their own independent research of anti-tax materials they obtained over the Internet or through seminars. This then could lead to the progressive adoption of additional far-right extremist belief. Others find opportunities to engage in financial crimes through their networks. These opportunities are based on relational connections to others, including co-offenders and victims, as well as family, friends, and 6 business associates. These relationships could contribute to participation in tax avoidance for financial gain, only to progressively lead to other, more ideologically motivated criminal behavior. Additionally, those in preexisting conflict with the Internal Revenue Service (IRS) over unpaid taxes may seek out ant-tax arguments to legitimize their desire not to pay their tax debts. Network connections then facilitate the ability to carry this out and could lead to engagement in other types of criminal behavior. This highlights the various opportunities and motivations involved in far-right extremist financial crime. Far-Right Extremists and Financial Crime Far-right extremists are involved in numerous financial and non-financial crimes, most typically tax avoidance/refusal. They utilize various strategies in these financial schemes, including “zero return”, corporation sole, offshore accounts, nominee entities, abuse of charitable deductions, and failure to file an income tax return.2 To date, little attention has been paid to farright extremist crime. However, crimes committed by extremist tax protesters should not be overlooked. Over 609 financial schemes have been committed by far-right extremists in the United States since 1990 (Sullivan, Freilich, & Chermak, forthcoming 2015b) with 445 of these involving identifiable tax protesters (Sullivan, Freilich, & Chermak, forthcoming 2015a). Over 700 individuals involved in the anti-tax movement have been indicted for financial crimes in the United States between 1990 and 2010, resulting in a conservative estimate of over one billion dollars in government losses (Sullivan et al., forthcoming 2015a). Financial crimes carried out by extremists disrupt and defraud both citizens and governments while simultaneously supporting violent extremist crimes in some cases, representing an area of growing concern. In particular, liens and other frivolous legal documents filed against public officials are frequently in direct retaliation for the enforcement of tax and 2 See IRS annual publication on the most popular tax fraud schemes, or the “Dirty Dozen” (IRS, 2013). 7 other laws by all levels of government. Tax avoidance and bank/check fraud typically stem from a desire to disrupt the U.S. financial system. Despite these dangers, far less attention from policymakers and scholars has been devoted toward far-right domestic extremism than to international extremism (i.e. jihadists) and the financial and organizational aspects of extremism in particular have not received adequate attention. Far-right extremists have a range of motivations for engaging in financial crimes, from purely ideological based on anti-tax beliefs to non-ideologically linked offenders motivated by profit or greed (Belli, 2011; Sullivan et al., forthcoming 2015b). These motives are not necessarily constant, as participation in these crimes varies drastically over time. Those who develop extremist ideologies and engage in ideologically motivated offending may exhibit drastically different behaviors over their life course. This could include either desistance from crime and denouncement of extremism to progression toward increasingly volatile, ideologically driven behavior. Others could move from holding anti-tax beliefs and engaging in crimes to maintaining a distrust of the government and hated of paying taxes, but behaving in ways that comply with the law, lessening their ideological stance to avoid further criminal sanctions. As noted above, relational connections to others involved in the same or similar far-right ideological movements are important influences on beliefs and behaviors. Despite a growing body of literature focusing on the financial activities of extremists generally (mostly focused on terrorism financing), few empirical studies have specifically examined extremist financial crime networks. Among the limited number of studies on extremist financial crime, network studies of far-right extremist financial crime are virtually nonexistent. 8 Network Analysis and Extremist Financial Crime Little research has been conducted on financial crime related to extremist movements, which is a substantial shortcoming in our current understanding of how extremists operate, organize, and engage with non-extremists in committing crime. This study takes a network analytical approach to examining the anti-tax movement in the United States. As opposed to traditional research methodologies that assume independent observations, network analysis accounts for interdependent connections among individuals. This approach is grounded in situational and relational theories of behavior by focusing on opportunities and motivations to commit crime in the complex, open system of the anti-tax movement. Network analytical techniques have tremendous potential to inform empirically derived enforcement and prevention policies and practices. “Systematically accumulating knowledge about the structural ‘blue print’ of criminal activity increases our understanding of their functioning and flaws, and may lead to effective ways to counteract and disrupt those networks” (van der Hulst, 2009, p. 102). Network analyses have been growing and yielding new insights in the study of fraud (Baker & Faulkner, 1993, 2003; Faulkner, Cheney, Fisher, & Baker, 2003), street gangs (McGloin, 2005a, 2005d; Nguyen & McGloin, forthcoming; Papachristos, 2006, 2009; Pettersson, 2003; Sarnecki, 2001; Sarnecki & Pettersson, 2001), illicit markets (Iwanski & Frank, 2014; Malm & Bichler, 2011; Malm, Kinney, & Pollard, 2008; Morselli, 2001; Morselli & Petit, 2007; Natarajan, 2006), and organized crime syndicates (Coles, 2001; Klerks, 2002; McIllwain, 1999; Morselli, 2003). Important terrorism research utilizing network analysis has focused mostly on Jihadi terrorist networks (Koschade, 2006; Krebs, 2002; Leuprecht & Hall, 2013; McAllister, 2004; Medina, 2014; Mullins & Dolnik, 2010; Sageman, 2004, 2008) as opposed to the extremist far-right (for expections, see Belli, 2011; Caspi, Freilich, & Chermak, 9 2012). In addition to examining Jihadi networks, far-right extremist networks are also important (Asal & Rethemeyer, 2006), particularly when addressing financial crimes. How network structures facilitate criminal action is currently not well understood, underscoring the importance of this research. Networks of financial crimes and perpetrators develop through cumulative local social interactions. Sometimes these interactions produce cohesive subgroups of individual offenders participating in common financial schemes together. Subgroups have the potential to cultivate knowledge of criminal techniques that increase the capacity for carrying out successful schemes (Burt, 2005; Frank & Zhao, 2005; Nonaka, 1994). This results in similar criminal behaviors among like-minded others. From a systems perspective, the emergence of relational connections leads to other connections of various types to increase the capacity for successfully carrying out criminal activities. Connections among subgroups, or bridging ties, are important links in the network. These are exploited to take advantage of new information and resources not available within subgroups, develop collaborative efforts among groups, or establish strategic links to new subgroups (Asal & Rethemeyer, 2006; Matthew & Shambaugh, 2005). Identifying the evolution of relational connections among offenders in the context of cohesive subgroup in addition to co-offending networks and links to larger organizations can further inform intervention strategies. Policy Implications The proposed study has clear implications for policy and practice. One of the key questions generating debate among scholars, practitioners, and policymakers centers on strategies to address extremist criminal activity. Many policies promote law enforcement responses centering on arrest and legal sanctions, but these have mixed outcomes depending on the nature of the group structure and different criminal activities undertaken over the life course 10 (Freilich, Chermak, & Caspi, 2009). The structure of extremist crime activity has largely shifted over the last two decades, moving from centralized control to lone offenders who identify with a general movement but act independently (Damphousse, 2010; FBI, 2007; Gruenewald, Chermak, & Freilich, 2013a, 2013d; Kaplan, 1997; Michael, 2012). Others are involved in network structures of varying compositions from more informal associations that facilitate the commission of isolated financial schemes to larger, more complex organizations. Focusing on these different network structures can illuminate new directions for law enforcement intervention the development of relevant public policies to address these crimes. Understanding the network dynamics through a problem-oriented approach is important for intelligence gathering and information sharing across agencies. Evidence-based knowledge of the causes and consequences of extremist financial crime can enhance the decision-making capabilities of policymakers and practitioners. This study will enhance the state of knowledge of these how these crimes are organized by focusing on network dynamics, contributing to both policy and practice. Research Project The goals of this study are to identify the composition, organization, and motivation of financial crime schemes and criminal offenders in the far-right anti-tax movement and examine how these networks change and evolve over time. The study utilizes a mixed method, multilevel, longitudinal network analysis of financial schemes linked to the American far-right antitax movement using data from the United States Extremist Financial Crime Database (hereafter referred to as EFCDB), a subsection of the U.S. Extremist Crime Database (ECDB).3 The 3 The U.S. Extremist Crime Database (ECDB) is funded by the National Consortium for the Study of Terrorism and Responses to Terrorism (START), a U.S. Department of Homeland Security (DHS) Center of Excellence. See Freilich et al. (2014) for complete description of ECDB development, coding criteria, and preliminary findings, Sullivan et al. (forthcoming 2015b) for financial crime schemes involving far-right extremists, and Sullivan, Freilich, and Chermak (2014) for financial crime and material support schemes linked to al Qa’ida and affiliated movements (AQAM). 11 network analyses in this study examine the relational dynamics among criminal suspects engaging in financial schemes. This is accomplished through several sequential analyses. First, a descriptive analysis examines several variables central to understanding extremist financial crime schemes and perpetrators, including scheme type, size, ideological affiliation, length, motivation, and demographics. Second, identifying cohesive subgroups determines the extent to which the network is clustered into suspects and schemes occupying the same social positions, where individuals are more likely to participate in the same schemes within their subgroup than outside of their subgroup. Third, a qualitative analysis of links between subgroups reveals important themes underlying the changing organization and composition of the network. Visual examinations of the network over time further illuminate the subgroup separations, links among subgroups, and how these change over time. Research Questions This study will address the following research questions. 1) What are the network connections among far-right anti-tax criminals and their collaborators in financial schemes? 2) To what extent are criminal networks in the far-right anti-tax movement organized in cohesive subgroups? 3) How do criminal networks in the far-right anti-tax movement change over time? 4) What characteristics of both financial schemes and criminal suspect shape network relational connections in the far-right anti-tax movement network? 12 Data, Methodology, and Analysis This study stems from the U.S. Extremist Crime Database (ECDB) study. The ECDB is an open source, relational database consisting violent incidents, financial schemes, and extremist groups. Violent incidents include homicides committed by extremists; bombings and arsons committed by environmental and animal rights extremists; and plots inspired by al Qa’ida and associated extremist movements (AQAM) (Freilich et al., 2014), while the group database includes domestic right-wing extremist groups (Chermak, Freilich, & Suttmoeller, 2013). As stated above, the data for the current study are derived from the EFCDB, which comprises all financial crimes committed by political and religious extremists from 1990 to 2010. The EFCDB includes rich qualitative case histories from open source data of crimes involving many extremist and non-extremist criminal perpetrators and hundreds of variables coded in a relational database capturing both quantitative and qualitative characteristics of financial schemes and criminal suspects. This study takes advantage of these qualities by combining the strengths of these approaches with social network analysis to examine the characteristics of how and why tax protesters organize to engage in financial crimes. This study utilizes a strategy of “open system” network construction, meaning every offender and scheme meeting the broadly outlined inclusion criteria is part of the study as part of the whole network, regardless of whether or not connections between components are present. In network analysis, rules of random sampling and independence of observations do not apply, as many network members are inherently connected in some way (Wasserman & Faust, 1994). Boundary considerations are established as follows: (1) criminal charges issued through a U.S. court of any jurisdiction (behavioral criterion); (2) the criminal activity took place in the U.S.; (3) the crime involved a financial scheme, such as tax fraud, money laundering, investment 13 fraud, check fraud, or false financial liens; (4), the crime occurred at some point between 1990 and 2010; and (5) all indicted suspects, including non-ideological collaborators, are included if at least one suspect adhered to a far-right extremist anti-tax ideology (attitudinal criterion). The ECDB relies on open source data collection methods to obtain information on financial schemes and is coded in a relational database (Freilich et al., 2014). A list of all publicly available cases was created using known databases, court documents, media records, and watch-group reports. Each case was treated as an individual case study for the purposes of identifying relevant schemes and suspects involved in each scheme. Official records (court, government documents), existing terrorism database, watchdog group reports, media reports, and websites were mined to gather all available information about each case (Freilich et al., 2014). Archival records are often used as a data source for network analysis, allowing the ability to track relationships in an unobtrusive manner when face-to-face interaction is either too difficult or not feasible (Alexander & Danowski, 1990). The relational database was used to code the data into distinct databases: (1) financial scheme (i.e. information on the criminal events and activities themselves) and (2) criminal suspect (i.e. known perpetrators involved in the scheme), an innovation extending beyond the most efforts focusing only on a single unit of analysis. Data were exported from EFCDB into various network data formats to conduct analyses in sequential steps: (1) cohesive subgroups, (2) network visualization, and (3) qualitative analysis. Examining the presence or absence of cohesive subgroups in far-right anti-tax networks addresses the second research question. Subgroups are important characteristics that uniquely account for aspects of social structure, behaviors, and attributes by focusing on persistent patterns of interaction (Frank & Yasumoto, 1998). This analysis identifies clusters of those who 14 engage in frequent interactions by participating in common schemes, the underlying premise being intra-cluster similarity and inter-cluster dissimilarity. Subgroup members exhibit a preference for co-offending together, offering structural advantages due to similar behavioral norms and resource sharing capabilities used to facilitate criminal activities. This study utilizes Frank’s (1995, 1996) KliqueFinder algorithm to determine the presence or absence of cohesive subgroups. The cluster assignments represent subgroups where suspects are more likely to cooffend in schemes together but not with those outside of their subgroup. In addition to the statistical analysis of the data for the presence or absence of subgroups, visual examinations of the network are conducted to illuminate different characteristics about network structure and composition. This addresses each of the four research questions through the visual examination of network, scheme, and suspect characteristics over time. Visualizations involve the creation of network graphs (sociograms) depicting relations among suspects and schemes in the network. Changes in connections among network members are examined by comparing separate network images identifying active subgroup memberships and connections each year for similarities and differences in network structures. Preserving the schemes as part of the social structure offers insight into specific offending patterns. The results of the statistical analysis for cohesive subgroups and network visualizations are further explored through an in-depth qualitative analysis of the networks, including connections between subgroups. This addresses the third and fourth research questions. EFCDB source documents are consulted to determine common themes explaining the dynamics of the network relations over time. This involved a detailed evaluation of the case histories to build explanations for the resulting subgroups and inter-subgroup connections over time. By combining qualitative analysis with the quantitative and graphic network approaches outlined 15 above, more complete and richer insights into the formation, evolution, and desistance of financial crime schemes is achieved. Overview of Subsequent Chapters This dissertation consists of the following chapters: Chapter 2 discusses relevant literature on far-right extremist financial crime, the American anti-tax movement, and the structure of their criminal activities. Chapter 3 reviews theories explaining anti-tax beliefs and corresponding criminal behaviors. Chapter 4 outlines methods for data collection, operationalization, and analysis. Chapter 5 provides a descriptive analysis of the data. Chapter 6 presents findings from the cohesive subgroup analysis. Chapter 7 examines these subgroups qualitatively, with a particular focus on associations to larger anti-tax organizations. Chapter 8 discusses implications and directions for future research. 16 CHAPTER 2: LITERATURE REVIEW This chapter provides the background for the study with an overview of the relevant literature on white-collar and organized crime, far-right extremist financial crime, the American anti-tax movement, and the structure of criminal activities involving far-right extremists. White-Collar and Organized Crime White-collar crime, organizational crime, organized crime, and the specialized area of extremist financial crime are particularly challenging to conceptualize and greatly overlap in many ways. Settling on a single classification would grossly misrepresent their unique characteristics. This study focuses generally on the organization of criminal behavior based on opportunities to commit various types of crime. This section examines research on various elements of organizational and financial crimes in order to situate the current study in the context of existing scholarship. In many cases, the lines between white-collar crime and organizational crime have been blurred, and sometimes the two cannot be distinguished (Albanese, 2000; Friedrichs, 2004). Defining the differences in these crime classifications has proven exceptionally difficult, as scholars and practitioners have failed to agree on common conceptualizations (J. Braithwaite, 2000; Geis, 2007; Griffin, 2002). The majority of academics in other fields have utilized the offender-based definition of white-collar crime first popularized by Sutherland (1940, 1944, 1949, 1983). However, not all offenders who carry out financial crime (typically categorized as white-collar crime) are of high socioeconomic status. Others fall somewhere between the high status offender and lower status street offender, including those with occupations more typical of the majority of the labor force. Indeed, prior research has found white-collar crime to be spread across social classes (Weisburd, Wheeler, Waring, & Bode, 1991). For instance, a plumber who 17 deliberately charges customers for expensive services that were not performed, were performed poorly, or could have been addressed by far simpler and cheaper means has committed fraud through deception and abuse of trust, but would not be a white-collar criminal according to Sutherland (Benson & Simpson, 2009). By focusing disproportionately on the status of the offender, harms caused by the actual offense can be missed. While it is important not to lose sight of the harm caused by offenders and offenses not typically captured by the criminal justice system, the problem of white-collar and organizational crime in particular is far more encompassing than high status offenders alone. Benson and Simpson (2009) have pointed directly to opportunities available to carry out different types of crimes as a way to reconcile the differences between offender and offense based conceptualizations of white-collar crime. Sutherland (1940, 1944, 1949, 1983) recognized that every occupation inherently contains unique relational dynamics that allow for different crimes to occur. These crimes revolve around the misuse of a position of power, causing a breach of trust held by a person in that position. That is, person must be in a position to commit a crime by deception, misrepresentation, or manipulation in order for a white-collar crime to occur (Benson & Simpson, 2009). Those in higher social classes are more likely to occupy these positions and should be more likely to commit white-collar offenses due to increased opportunity (Benson & Simpson, 2009). Lower socioeconomic status individuals who do not occupy higher positions of trust do not have the same opportunities to commit white-collar offenses, and therefore are more likely to commit street-level offenses (Benson & Simpson, 2009). In particular, criminals take advantage of trust and use their unique position to perpetrate different types of financial frauds. 18 Financial crime is related to organized crime, which is traditionally framed in terms of profit-driven criminality involving multiple actors working together in some organized manner to commit criminal acts. Financial crime is intricately involved in the operation of criminal enterprises (Beare, 2003; Levi, 2003). However, like white-collar crime, organized crime (or criminal organization) is difficult to conceptualize (van der Beken, 2004). Albanese (2000, p. 411) provided a definition for organized crime combining definitions from numerous authors: “organized crime is a continuing criminal enterprise that rationally works to profit from illicit activities; its continuing existence is maintained through the use of force, threats, monopoly control, and/or corruption of public officials.” This is one of many definitions in the literature, but captures numerous elements typically considered part of organized crime. While organized crime is often considered in terms of hierarchical criminal organizations, this view is largely incomplete and misleading. It can range from these larger, politically integrated crime organizations to malleable but interconnected networks of individuals, some of which are more heavily involved in crime than others (Albanese, 2011). They engage in business enterprises together, exploiting opportunity for mutual benefit. Given these substantial differences in what constitutes organized crime, organized crime can be thought of as one type of crime under the broader category of organizational crime, which focuses on the organizational aspects of how criminals behave collectively in different contexts (Albanese, 2000). Organizational components are important influences on the commission of financial crime. Organizational crimes are typically conducted to enhance the interests of a company, agency, or group as opposed to the exclusive benefit of the individual (Gross, 1980; Reiss & Tonry, 1993). Prior research has found evidence of an exchange between so-called legitimate and illegitimate business, with numerous scholars arguing that both should be examined 19 conjointly (Albanese, 2011; Middleton & Levi, 2005; Morselli & Giguere, 2006; Passas, 2003; Ruggiero, 2003; van Duyne, 1993). Organized criminal elements become involved in legitimate business whose cooperation is necessary to carry out the crimes. Such is the case with money laundering, which has been a key element in the successful operation of criminal enterprise, particularly for terrorism financing (Shelley, 2014). Many high-profile cases bear this out, including the recent settlement with HSBC Bank over laundering of drug money and terrorist money (U.S. DOJ, 2012). Fitting with the theme of challenges brought about by trying to properly classify organized and white-collar crimes, financial crimes by far-right extremist tax protesters defy classification in many ways. While many are committed as part of their regular occupational duties, others engage in separate criminal schemes not tied to their employment. Crimes are committed by both loosely organized networks of offenders with different motives, skills, and techniques, but also by individual offenders acting on their own behalf and not directly tied to others. Tax avoidance by ideologically motivated individuals may or may not be related to financial need or want (either individual or organizational), as most other types of financially related crimes are to some extent. However, the attention given to ideologically motivated crimes by extremists has been minimal. This study takes a step toward addressing this gap. Far-Right Extremist Financial Crime Far-right extremist financial criminals have not been examined in the literature or received adequate attention from policymakers and practitioners. These individuals base their criminal activities on an ideological belief system outside of generally accepted mainstream political and religious views, such as terrorists and extremists. While financial crime research has traditionally been centered more generally on affluent and relatively powerful white-collar 20 criminals, financial crime is attractive in all areas of society. This growing body of research has examined the crime-terror nexus, or the connections between criminal activities and ideologically motivated extremist groups. While motivations for extremist financial crimes can sometimes be easy to ascertain, others are not as straightforward (Belli, 2011). There are various motives for crimes involving extremists, some of which are conducted to finance extremist organizations generally, such as through money laundering and material support for terrorism (supplies, training, personnel, etc.) (Sullivan, Freilich, et al., 2014). Others have ideological motives for the specific financial crime acts themselves, such as extremist offenders who utilize various frivolous arguments and conspiracy theories to justify their actions. These offenders may participate in fraud for financial reasons, but also maintain a belief system that, at least in part, drives their behavior. For example, many extremists try to disrupt the U.S. financial system or believe the laws of U.S. do not apply to them as independent ‘sovereign citizens.’ Many act on these beliefs by engaging in financial crime schemes, such as tax avoidance, identity theft, counterfeit financial instruments, false liens, and various other frauds. Extremists commit a wide array of financially related (non-violent) offenses, including tax fraud, money laundering and dirtying, identity theft, and counterfeiting (Freilich et al., 2014). However, empirical attention to the intersections of far-right extremism and financial crime has been minimal. The literature tends to utilize anecdotal evidence and single-source case studies (Gruenewald et al., 2009). The literature on terrorism financing generally has been far more extensive than studies examining the intersections of far-right extremism and financial crime. Despite this shortcoming, far-right extremists in the U.S. continue to be heavily involved in both financial and violent crimes. This is a notable oversight, as far-right extremists in the U.S. have 21 committed over 609 financial schemes from 1990 to 2013 (Sullivan et al., forthcoming 2015b). There are exceptions, including Belli’s (2011) study of financial crimes committed by networks of far-right and jihadi extremists. However, little remains known about far-right extremist financial crime. The current study adds to the existing literature by mapping the network connections among far-right extremists involved in financial crimes. More specifically, it focuses on the antitax movement, which exhibits characteristics similar to but distinct from other types of general far-right ideological groupings (e.g. white-supremacists, militias, sovereign citizens, etc.). In order to properly specify the scope of this study, a general description of the history of the antitax movement, its belief system, the tactics and strategies used by tax protesters, and its connection to the sovereign citizen movement are necessary. The next section describes these characteristics of the American anti-tax movement in greater detail. American Anti-Tax Movement For the purposes of this study, anti-tax protesters are considered a subsection under the general umbrella of far-right extremist ideology. The anti-tax movement is one of the longest running anti-government movements in the United States, with roots dating back to tax protests precipitating the American Revolution. This anti-tax sentiment has persisted throughout the history of the United States, but the modern anti-tax extremist movement developed in the 1950s along with increased anti-government activism through an extensive array of beliefs regarding the legality of the tax system. Tax protesters believe they have a moral or legal obligation to refuse to pay taxes and follow other laws considered illegal or unconstitutional. Beginning with more mainstream attempts to repeal the Sixteenth Amendment to the Constitution authorizing a federal income tax, tax protesters eventually developed various alternative theories of the 22 amendment and the tax code, using these misinterpretations as legal justification for tax avoidance (ADL, 2005). One of the early adopters of anti-tax philosophy was Arthur Porth, who in the early 1950s began utilizing a number of different strategies later adopted, altered, and expanded by other members of the movement (ADL, 2005; Levitas, 2001). These included the Sixteenth Amendment being unconstitutional, the Thirteenth Amendment barring involuntary servitude, the Fifth Amendment against self-incrimination, and gold and silver not backing Federal Reserve Notes (paper money) disqualified them as currency, thus making them untaxable (see IRS, 2014, outlined in further detail below). He was also one of the first to distribute anti-tax literature, a tactic that remains popular among tax protesters (ADL, 2005; Levitas, 2001). He was eventually convicted of tax evasion. While certainly not the first tax protester, his and other similar stands taken against taxation spawned a growth movement in the decades that followed (ADL, 2005; Levitas, 2001). In fact, numerous anti-government extremists, such as William Potter Gale, founder of Christian Identity group Posse Comitatus, and Martin A. Larson, contributing editor of far-right magazine American Mercury, praised Porth, imitated his tactics, and promoted his ideas (Levitas, 2001). Other prominent tax protesters would follow. One was Irwin Schiff, who promoted antitax ideas to a more mainstream audience, appearing on numerous television and radio shows since the mid 1970s (ADL, 2005; Levitas, 2001). He has been convicted of tax avoidance numerous times and was a prominent promoter of anti-tax literature (Levitas, 2001). Martin Cooley was another tax protester who conducted seminars beginning in the 1960s and wrote a popular anti-tax book complete with sample letters and tax documents (ADL, 2005; Levitas, 2001). He would eventually serve multiple prison terms for tax evasion. It is notable that one of 23 those who attended Cooley’s seminars and followed his anti-tax strategies was Robert Jay Matthews, who would later start the neo-Nazi group “The Order” (ADL, 2005; Levitas, 2001). Other tax protesters from the 1970s and 80s include former Illinois Department of Revenue agent William Benson, Montana anti-Semite Martin Beckman, and Posse Comitatus member Gordon Wendell Kahl, who shot and killed two U.S. marshals in North Dakota and a sheriff in Alabama in 1983 (ADL, 2005; Levitas, 2001). In light of the substantial growth in the anti-tax movement by the early 1980s, the government took steps to address the problem with increased prosecution and penalties for tax avoidance (Levitas, 2001). However, the Congress in the late 1990s limited the ability of the IRS to address tax avoidance by cutting the IRS budget amid hearings consisting of much of the same anti-tax rhetoric seen in extreme far-right arenas over the previous three decades (Levitas, 2001). Among their actions was barring the use of the term “tax protester” by IRS agents to describe taxpayers using traditional anti-tax rhetoric (ADL, 2005). A new wave of tax protests and tax avoidance would follow as audits and criminal and civil actions against tax evaders substantially declined throughout the 1990s (ADL, 2005). This resulted in a combination of increased motivation and rationalization to engage in tax avoidance coupled with substantially lower risks of detection (Friedrichs, 2004). Several notable organized anti-tax groups gained traction in the early 2000s. One was the “We the People Foundation” led by Bob Schulz, which promoted itself as an education program conducting seminars across the country and even purchasing full page advertisements in USA Today (SPLC, 2001). Other organizations included “We the People” based in California and the Save-a-Patriot Fellowship (ADL, 2005). Many people publicly announced they would no longer pay taxes or withhold taxes from their employees. Although the resurgence of this anti-tax 24 sentiment vastly increasing tax avoidance has subsided somewhat in recent years, the wide-scale anti-tax movement persists in the present day. Frivolous Tax Arguments As noted above, the most prevalent far-right extremist financial crime activity is carried out by the anti-tax movement, whose proponents utilize various “frivolous arguments” to provide a moral or legal justification for not following tax laws. They follow an ideological motivation for anti-tax behavior, which becomes obligatory to its true believers. According to the Internal Revenue Service (2014), frivolous tax arguments fit under five broad categories: 1) voluntary nature of the federal income tax system, 2) meaning of taxable income, 3) meaning of terms used in Internal Revenue Code, 4) claims regarding constitutional amendments, and 5) fictional legal arguments. These arguments are summarized below (see IRS, 2014, "The Truth about Frivolous Arguments" for a complete overview). Those who argue that the federal income tax system is voluntary do not believe they are required to pay taxes or file tax returns, pointing to the reference to a “voluntary” system in the Form 1040 instruction book, which refers to the ability of taxpayers to determine the correct amount of tax owed when filing their tax returns. When this fact is pointed out, proponents of the voluntary law argument will demand the IRS identify the relevant laws imposing taxes, then dismiss any attempts by the IRS to fulfill this request. Others argue that failure to respond to their demands within a certain time frame constitutes acceptance of their claims and negates any requirement from the individual to pay taxes owed. A different variation of this centers on the argument that administrative summons issued by the IRS are also voluntary. Other tax protesters believe that the IRS is required to prepare tax returns for those who do not file their own returns under section 6020(b). Another method is known as “zero return,” where the tax protester will 25 file a tax return claiming no income and no tax liability. These are often filed as amended returns asking for refunds of previously paid taxes. The second common approach taken by tax protesters centers on the meaning of income, where proponents argue that their income does not fall under the definitions of taxable income, thereby relieving them of the requirement to pay taxes or file tax returns. These arguments include: • Wages, tips, and other compensation received for personal services are not taxable. • Only foreign source income is taxable. • Federal Reserve Notes are not income. • Military retirement pay does not constitute income. The wages not constituting income argument essentially asserts that money in exchange for labor or time provided for service does not constitute a gain or profit, therefore no tax should be administered, as the Sixteenth Amendment to the Constitution only authorized taxation of profit. The foreign income is a variation of the income arguments, stating that income taxes can only be assessed against nonresident aliens and foreign corporations. The third argument is that Federal Reserve Notes are not currency that can be exchanged for gold or silver and therefore cannot be taxed. Despite these arguments, the courts have consistently held that all income is taxable, regardless of its source. A third area of frivolous tax arguments centers on misinterpretations of various terms used in the Internal Revenue Code. The most prevalent of these is the argument that the individual is not a “citizen” or “person” and therefore not subject to federal income tax laws. These arguments are typically used by sovereign citizens who either reject U.S. citizenship and claim state citizenship or deny citizenship (personhood) entirely. A variation is the claim that the 26 “United States” excludes the “sovereign” states and only includes the District of Columbia and federal territories and enclaves. Those not in the “United States” are not subject to federal laws, including taxes. Another argument is based on a misinterpretation of section 3401, which governs taxation of government employees, stating that only these employees are subject to federal income tax law. Another type of frivolous tax argument deals with claims regarding constitutional amendments. These claims are various and therefore stated here without further explication: • Taxpayers can refuse to pay income taxes on religious or moral grounds by invoking the First Amendment. • Federal income taxes constitute a “taking” of property without due process of law, violating the Fifth Amendment. • Taxpayers do not have to file returns or provide financial information because of the protection against self-incrimination found in the Fifth Amendment. • Compelled compliance with the federal income tax laws is a form of servitude in violation of the Thirteenth Amendment. • The federal income tax laws are unconstitutional because the Sixteenth Amendment to the United States Constitution was not properly ratified. • The Sixteenth Amendment does not authorize direct non-apportioned federal income tax on U.S. citizens. The final category of frivolous arguments includes legal fictions. The first argues that the IRS is a private corporation, not a U.S. government agency, and therefore has no legal authority to enforce the Internal Revenue Code. Another argument is that the Paperwork Reduction Act of 1980 provision requiring an Office of Management and Budget (OMB) control number on all 27 requests for information invalids penalties imposed for not filing a tax return because the instructions for Form 1040 do not have the OMB number. A third argument is that African Americans can claim a special tax credit as slavery reparations. Variations of this contention are promoted by those associated with the “Black Nationalist” or “Moorish Nationalist” movement. Others utilize fictitious legal techniques. “Untaxing” packages or trusts often sold by those who incorporate other frivolous arguments and claim these instruments remove a person from federal taxation requirements. The “corporation sole” scheme involves illegally applying for incorporation as a religious entity exempt from taxation. Whereas a true corporation sole allows religious leaders to conduct business (contracts, property purchases, etc.) in their official capacity but not for personal benefit, tax protesters use the technique to avoid lawful taxation. Another popular technique among sovereign citizens is based on the “redemption” or “strawman” conspiracy theory. It states that the U.S. Treasury Department creates secret bank accounts for each U.S. citizen called a “strawman” that can be utilized by those citizens to pay taxes and other debts through the use of fake financial instruments (often fictitious Treasury checks known as a “sight draft” or “bill of exchange”).4 The theory also claims that Form 1099OID can be sent to creditors, who can present them to the Treasury to satisfy personal debts. Other arguments are often made regarding the tax collection process. These too are numerous and will not be replicated here, but typically revolve around some issues with the type of paperwork, instruction, or form. Some claim these do not have the proper signatures, such as from the Secretary of the Treasury. Others argue the proper paperwork authorizing government action was not produced or the incorrect form was issued (IRS, 2014). Another line of dispute challenges the legal authority of the tax law, tax court, or IRS agents. These challenges in court 4 For more information on sight drafts, bills of exchange, or the redemption/strawman conspiracy theory, see U.S. Department of Treasury (2014), Southern Poverty Law Center (2005, 2010), and Sanchez (2009). 28 are sometimes accompanied by legal actions taken against individual agents, other officers, or judges, such as the filing of false liens, tax forms, or other legal documents against the officials. This course of action is popular among sovereign citizens, who are closely aligned with the antitax movement. Connections to Sovereign Citizen Movement Anti-tax arguments that the tax system and tax laws are illegal and unconstitutional often flow easily into the belief that the entire government is corrupt and illegitimate, which can lead to the adoption of other extremist ideologies, such as those of sovereign citizens, militia and Patriotism, and Christian Identity (ADL, 2005). In particular, a large number of those who utilize frivolous tax arguments and believe in their moral or legal duty to oppose taxation also adhere to a “sovereign citizen” philosophy, which argues that the individual is either a sovereign entity not subject to government authority or a citizen of a fictitious state body, such as a “sovereign” state republic. Sovereign citizens adhere to their own version of the law, which is a convoluted combination of the Magna Carta, the Bible, English common law, and out of date constitutional law. They often use punctuation when spelling their name, especially to separate the first and middle names from the “government issued” last name (ADL, 2012). Sovereign citizens believe the legitimate government was long ago replaced with an illegitimate government that entered into secret contracts with citizens through various government requirements (e.g. driver’s license, zip code) to take away their sovereignty. By renouncing their allegiance to this illegitimate government, citizens can reclaim their sovereignty and remove any authority the government holds over them (ADL, 2012). There is tremendous overlap between sovereign citizens and tax protesters. Sovereign citizens are also closely aligned with other far-right extremist ideologies, including the Patriot 29 movements, anti-government militias, and white supremacists, and are also referred to as “freemen” or “constitutionalists.” Although the number of sovereign citizens is impossible to know due to their individualistic and secretive nature, the numbers are estimated to be in the tens of thousands (ADL, 2012). Much of the sovereign citizen philosophy stems from the anti-government, anti-Semitic Posse Comitatus group from the 1970s, who originated the concept of “common law courts” to issue arrest warrants and subpoenas to judges, law enforcement, and other officials (ADL, 2012; Sanchez, 2009). Sovereign citizens have utilized these types of courts and frivolous legal documents stemming from them frequently over the last four decades. Sovereign citizen beliefs are spread largely online, through seminars and literature, and between prisoners (ADL, 2012), in much the same way as anti-tax philosophies and techniques. The belief that the government holds no legal authority over them puts sovereign citizens in frequent and persistent conflict with governmental authorities of all levels and jurisdictions, including law enforcement. They commonly employ techniques characteristic of sovereign citizens including “paper terrorism,” which includes filing false liens or other legal documents against officials and others who have “wronged” the sovereign individual or others in the movement (ADL, 2012; Barkun, 1997; Flynn & Gerhardt, 1995; Pitcavage, 1998, 1999). There is further evidence of an “escalation” in radicalization, where individuals start extremist activity by engaging in ideologically motivated tax avoidance or other financial schemes but later engage in more violent activity as their commitment to extremist movement increases (Freilich & Chermak, 2009). Sovereign citizens are frequently involved in violent altercations with law enforcement, posing a persistent threat to public safety. In a survey of local law enforcement, Carter, Chermak, Carter, and Drew (2014) found sovereign citizens to be the most serious threat 30 in their jurisdiction in 2013, up from the seventh most serious threat in a previous 2006 survey conducted by Freilich, Chermak, and Simone (2009). This threat is corroborated by the 145 ideologically motivated homicides involving 348 victims (including 32 law enforcement officers) carried out by far-right extremists (many of whom were sovereign citizens or subscribed to similar ideologies) in the United States between 1990 and 2010 (Freilich, Chermak, Gruenewald, & Parkin, 2012). Since 2010, violent criminal activity involving sovereign citizens has continued throughout the United States (ADL, 2012). Like the anti-tax movement, the “default structure of the sovereign citizen movement is that of a large mass of individuals or loosely aligned and informal/ad hoc groups, led by a number of sovereign citizen ‘gurus,’ who provide leadership and inspiration as well as new sovereign citizen ideas and tactics” (ADL, 2012, p. 6). Some larger sovereign groups have been developed, such as the Guardians of the Free Republic and the Republics for the united States of America (RuSA), but these are much less prominent than smaller, less organized, decentralized networks of individual sovereign citizens (ADL, 2012). Those who develop new sovereign legal theories are especially influential, as their ideas are spread to others for use throughout the United States. They utilize these various sovereign citizen and anti-tax arguments to engage in various types of criminal activities, both individually and as part of larger groups. The next section explores the structural patterns central to understanding of how financial crimes involving the anti-tax movement develop and function. Structure of Criminal Activities by Far-Right Extremists Far-right anti-tax protesters are involved in criminal activities in different ways. Some engage in singular schemes on their own without identifiable connections to others, which is often found with individual tax avoidance or filing false liens. Another common patterns 31 includes married couple engaging in schemes together, particularly tax avoidance (e.g. failure to file joint income tax returns). Others participate in schemes with informal networks of cooffenders, some of which form cohesive subgroups, while some are linked larger anti-tax organizations. This section examines both single offender (“lone-wolf”) and co-offending financial crimes by far-right anti-tax extremists. Lone Wolves The structures of extremist crime activities have largely shifted over the last two decades, moving from centralized control to “leaderless resistance”5 and “lone wolf” offenders who identify with a general movement and promote its goals but act independently of any formal group (Damphousse, 2010; FBI, 2007; Gruenewald et al., 2013a, 2013d; Kaplan, 1997; Michael, 2012). While lone wolf offenders are fairly rare in the United States, they are responsible for a disproportionate number of violent extremist incidents (Smith, Roberts, Gruenewald, & Klein, 2015). While studies of lone wolf financial crime offenders are limited, much can be learned from the growing literature on lone-wolf terrorism and extremism that can extend to individual offenders in explaining both the beliefs and behaviors necessary for participation in financial crime schemes. It is important to reiterate that lone wolf offenders may not be entirely disconnected, but have ties to an anti-tax ideological movement, sympathizing and identifying with far-right extremist beliefs (Gruenewald et al., 2013a). This does not necessarily imply direct connections or interactions with others, but could involve “self-radicalization” by obtaining information from indirect sources, such as the Internet or other literature (Gill, Lee, Rethemeyer, Horgan, & Asal, 2014; Gruenewald et al., 2013a). Those seemingly acting alone likely have some loose association to the broader movement, where ideas driving anti-tax ideology (including the wide 5 See Louis Beam’s 1992 essay “Leaderless Resistance” in the Christian Identity publication The Seditionist. 32 variety of common frivolous anti-tax arguments outlined above) are developed, engrained, and sustained (Moskalenko & McCauley, 2011). This promotes an “us” versus “them” mentality with the dehumanization of others in order to reduce psychological inhibitions toward violence (McCauley & Moskalenko, 2008; Spaaij, 2010). While some join extremist groups and are further radicalized into violence due to their increased associations with like-minded peers, others may be encouraged to take up violence on their own without joining any formal or informal group (McCauley & Moskalenko, 2008). The threat of far-right lone wolf extremism has received increased attention and is considered highly important to policymakers and practitioners. Spaaij (2010) found the most prominent ideologies involved in lone wolf terrorism in the United States to be white supremacist, nationalism/separatism, and anti-abortion (along with jihadist), although forty percent of the motivations were unknown. In a survey of state police agencies, Chermak, Freilich, and Simone (2010) found that fifty percent of respondents identified anti-tax protesters committing lone offenses as opposed to over twenty percent of Christian Identity proponents and thirty percent of sovereign citizens. This finding is logical given the nature of anti-tax crimes often consisting of individual tax avoidance. The prominence of lone wolf offending does not necessarily mean that far-right lone wolf extremism is increasing consistently over time. Using RAND Memorial Institute for the Prevention of Terrorism (MIPT) incident data, Spaaij (2010) found that lone wolf attacks increased sharply starting in the 1970s, peaked in the 1990s, and declined in the 2000s but to levels still above those of the 1980s. Using ECDB data, Gruenewald, Chermak, and Freilich (2013d) found further support for this decrease, noting a sharp downward trend in far-right lone wolf homicides after 2001 from the highs of the 1990s. This is in sharp contrast to the claims of 33 policymakers and those in the media who consistently claim rising levels of lone wolf terrorism. However, the large number of foiled and failed violent plots by far-rightists is not often considered in the same threat evaluations as homicides (Dahl, 2011) and may potentially offset the decrease in violent incidents with an increased number of unsuccessful lone wolf plots (Gruenewald et al., 2013d). It is also possible that other criminal activities by far-right lone wolves, including financial crimes, increased or remained steady during this same period. This question has not been subject to systematic study to date, although the current study takes the first steps toward addressing this knowledge gap. Co-Offending Networks In additional to lone-wolf crimes, far-right extremists co-offend in financial crime schemes with others as part of interconnected networks. Co-offending is central to understanding how individuals interact together to carry out financial crime schemes. It is also important to understanding how criminal networks function more generally. As Waring (2002, p. 43) has argued, “the treatment of co-offending as a network phenomenon is particularly powerful because it has the potential to contribute both to the incorporation of co-offending into models of other aspects of crime and to the location of social organization of crime within the broader range of forms of social organization.” Papachristos (2011) has also made a solid argument for the theoretical and methodological advancement of criminological research by incorporating the role of networks through a focus on interdependent interactions between social actors. These cooffending collaborations need to be considered at all stages of the criminal event to understand how criminal organizing is occurring (Tremblay, 1993). Financial crimes consist of various co-offending patterns, or shared connections between individual offenders in the context of carrying out financial crime schemes. Scholars have begun 34 to examine criminal enterprise in terms of dynamic networks as opposed to centralized, hierarchical structures (Coles, 2001; McGloin & Nguyen, 2014). While some are large organized groups, the majority consists of small number of individuals working together to take advantage of short-term opportunities (Levi, 2008). Those motivated by the same opportunity may decide to collaborate on a specific financial scheme before ending their criminal relationship and moving on to other criminal or non-criminal activities. Co-offending relationships are typically restricted to a small number of criminal events (Reiss & Farrington, 1991; Sarnecki, 2001), as criminal relationships are transient in nature (McGloin, Sullivan, Piquero, & Bacon, 2008). Individual offenders may also be involved in multiple criminal schemes simultaneously with different co-offenders. This complicates the study of the dynamics of co-offender relationships in the context of criminal events. Despite this propensity for co-offending in networks, this knowledge is not used to any great extent to determine intervention, prevention, and punishment (Kennedy, 2009). Cooffending criminal networks require their own theories and explanations, as they are distinctly different from other types of social networks (Morselli, 2009). Furthermore, the unique characteristics of co-offending behaviors may have their own distinct implications (Andresen & Felson, 2010). Targeting single individuals has limited effectiveness on groups of co-offending individuals (Andresen & Felson, 2010). Networks determine access to this specialized knowledge and resources necessary for carrying out crimes (constructing crime events) (Ekblom & Tilley, 2000). For example, a financial scheme based on the “redemption theory” and focused on selling debt elimination packages may require administrators to handle organizing personnel, (alleged) legal experts with knowledge of the tax code, seminar leaders to tell potential customers/victims about the theory behind scheme, and salespersons to get the product directly 35 to the customer/victim. Other key roles are necessary to carry out different types of criminal events. They also require the presence of those prone to believing the theories behind the scheme or with a desire for personal gain. These individuals provide the situational opportunity for criminals to carry out a financial scheme. This study utilizes network analysis to assess the relational connections of suspects cooffending in financial schemes. Network analysis is gaining more attention as a way to examine interconnections of criminal behavior (Carrington, 2011; Malm, Bichler, & Van De Walle, 2010; McGloin & Kirk, 2010a, 2010b; McGloin & Nguyen, 2012; McGloin & O'Neill Shermer, 2009; McGloin & Piquero, 2010; Morselli & Giguere, 2006; Morselli, Giguere, & Petit, 2007; Morselli & Roy, 2008; Papachristos, 2011; Papachristos, Meares, & Fagan, 2012; van der Hulst, 2009; Weerman, 2011). Several studies have focused on covert networks, such as street gangs (McGloin, 2005a, 2005d; Papachristos, 2006, 2009; Pettersson, 2003; Sarnecki, 2001; Sarnecki & Pettersson, 2001), illicit markets (Malm & Bichler, 2011; Malm et al., 2008; Morselli, 2001; Morselli & Petit, 2007; Natarajan, 2006), organized crime syndicates (Coles, 2001; Klerks, 2002; McIllwain, 1999; Morselli, 2003), and terrorist groups (Caspi et al., 2012; Koschade, 2006; Krebs, 2002; Pedahzur & Perliger, 2006; Perliger & Pedahzur, 2011; Sageman, 2004, 2008; van der Hulst, 2011). Despite these recent advances, network studies to date in the criminology and criminal justice literature have failed to focus on financial crime networks. Notable exceptions include network studies on price-fixing conspiracies in the heavy machinery industry (Baker & Faulkner, 1993; Faulkner et al., 2003) and Belli’s (2011) network study on far-right and Jihad extremist financial crime networks. Baker and Faulkner (1993) found network centrality to be key in facilitating communication and secrecy in the network, where business networks seek to 36 maximize efficiency while conspiracies instead seek to maximize secrecy. Faulkner and Baker (2003) found price-fixing cartels to be more effective through centralization of authority and continuity of actors. As noted above, numerous other studies have examined profit-drive criminal enterprises (organized syndicates, trafficking of drugs and other goods, etc.), but did not focus specifically on financial (white-collar) schemes. Much like the current study, Belli (2011) relied on the U.S. Extremist Financial Crime Database (EFCDB) to examine multiplex networks of far-right and Jihadi extremists, including criminal, business, and family ties. Belli (2011) utilized exponential random graph (also known as p*) modeling to examine various structural characteristics of these different networks. She found that by including informal network connections, subgroups tended to form around seemingly isolated components. More isolated components existed for far-right extremist networks than for jihadists. There were no substantial differences in the structure of far-right and Jihad co-offending networks, although business networks were much larger than criminal networks. This research further illuminated the value of examining multiple types of ties, different types of offenders (extremist and non-extremist), and informal group dynamics. While several network studies have examined co-offending and the two studies noted above have specifically focused on financial crime, no network study to date has accounted for the unique role of the criminal event in shaping co-offending behaviors. This study incorporates these different elements into the same analysis to provide a more complete picture of how individual, situational, and group-based elements are shaped in criminal networks. Conclusion This chapter provided an overview of the main areas of literature pertaining to the current study. The section on white-collar and organized crime highlighted the unique aspects of 37 extremist financial crime schemes within the scope of the existing literature. An overview of farright extremist financial crime, including the anti-tax movement, frivolous tax arguments, and the sovereign citizen movement, were essential to framing the scope of the research. The section on the structure of far-right extremist crimes reviews literature on both lone offenders and cooffending with a specific focus on network studies. The next chapter explores the theoretical basis for this study. 38 CHAPTER 3: THEORY This chapter focuses on theories explaining far-right extremist financial crime in the context of the anti-tax movement. Given the multi-dimensional and multi-causal nature of antitax offending, the review of theory is selective and purposive in order to avoid restricting the scope of potential theoretical explanations too narrowly. Theories from different but related disciplines shape how individuals engage with far-right extremist movements, adopt extremist beliefs and values, and ultimately interact with others to carry out financial schemes. Some are sympathetic to anti-tax views but do not act on those beliefs, while still others will act on their own without any real connection to others. Strong relational dynamics with similar others should reinforce both belief and behavior and increase the capacity for carrying out different types of crime. It is important to note that these theories are not being tested but instead used to enhance the understanding of how financial crimes related to the anti-tax movement develop, function, and evolve. Development of Extremist Beliefs Theories on the development of extremist beliefs generally stem from social psychology and sociology, but numerous criminological theories are also applicable in this context. McCauley and Moskalenko’s (2008, 2011) twelve mechanisms of radicalization include factors ranging from personal political grievance or victimization to feelings of strain and oppression from an external force threatening their values or identity (McCauley & Moskalenko, 2008, 2011). "Radicalization of many kinds may be associated with a syndrome of beliefs about the current situation and its history: We are a special or chosen group (superiority) who have been unfairly treated and betrayed (injustice), no one else cares about us or will help us (distrust), and the situation is dire—our group and our cause are in danger of extinction (vulnerability)" 39 (McCauley & Moskalenko, 2008, p. 416). Behavior on the other hand is associated with increased time, money, and risk spent in extremist activities, including joining extremist groups (McCauley & Moskalenko, 2008, 2011). This section focuses on the development of beliefs as opposed to criminal behavior, although intersections between the two are also discussed as they are inherently intertwined to some extent. The first mechanism discussed by McCauley and Moskalenko (2008, 2011) includes feelings associated with a personal grievance or victimization, where an individual perceives an injustice has been done to them or someone they know. In this view, the individual adopts extremist views or joins an extremist group to gain a measure of vengeance against their oppressors. This is consistent with general strain theory that argues that blocked opportunities, the imposition of negative stimulus, or the taking away of positive stimulus create grievances that lead to criminal behavior (Agnew, 1992, 2010). It is also consistent with resource mobilization theory, which focuses on individual frustration and anger when an individual focuses the blame for their existing circumstances toward an external reference group (Freilich, 2003). This can be found as a reason for the actions of many anti-tax extremists who believe that the IRS, or an IRS agent, has wronged them or treated them unfairly. This feeling can manifest itself in more general hostility and potentially lead to actions against the perceived wrongdoer, such as filing false liens or other legal documents or acting the anger out with violence. It can also be used as a catalyst toward the adoption of increasingly extremist beliefs and engagement with others holding similar views. Another mechanism is the identification with grievances shared by a larger group of people. Here, the focus is not on any personal identifiable incident, but a more general feeling of oppression, although it can coincide with individual victimization (McCauley & Moskalenko, 40 2008, 2011). Theories often focus on the perceived decline in status, such as structural changes in society that lead to the shifts in status away from once prominent groups, resulting in a backlash in order to deal with the resulting strain and anxiety (Aho, 1990; Hofstadter, 1964; Lipset & Raab, 1970). This includes the feeling that the broader group they identify with had been humiliated in some way, thus inspiring their adoption of extremist beliefs. "Positive identification with a group, combined with the perception that this group is being victimized, produces negative identification with the group perpetrating the injustice" (Moskalenko & McCauley, 2011, p. 122). This was found in several ethnographic studies of white supremacists in the United States (Ezekiel, 1995, 2002; Hamm, 1994). In the anti-tax movement, the oppressor is an illegitimate government authority that imposes illegal tax and other laws on American citizens. These individuals are likely to be part of a broader social movement with commonly held beliefs (see below for theories of participation in social movements). Other key radicalization mechanisms focus on the influences of other people on individual belief and behavior, both generally and in the context of extremist groups (addressed further in the next section). A powerful factor in determining individual adoption of extremist ideology includes relational connections, particularly family and friends. Numerous predominant criminological theories have recognized the role of deviant peer influences on criminal behavior, including social learning (Akers, 1973, 2011), differential association (Sutherland, 1940; Sutherland, Cressey, & Luckenbill, 1995), and social bond (Hirschi, 1969) theories. Connections can be based on a variety of potential ties, including friendship, kinship, or business ties, as persons with existing ties are socialized into a particular lifestyle of acceptance of crime or radicalization to an extremist ideology (Sageman, 2004). Sageman (2004) found that peer relations heavily influenced the decision to join Salafi jihad extremist groups and individuals 41 tended to join in groups rather than individually. These friendships preceded extremist beliefs, which only later developed through their interconnections with others first introduced through peer groups. They did not develop these beliefs on their own, but were taught by others. Bakker (2006) replicated Sageman’s study with a different set of offenders and found similar results. McCauley and Moskalenko (2008, 2011) also focused on connections to friends, family members, and lovers, where they were motivated to eventually embrace the extreme beliefs of those close to them. McCauley and Moskalenko (2011) also point to the mechanism of “unfreezing,” important life events (like the death of a loved one, or the ending of close friendships or marriages) that alter a person’s social connections leading them to embrace new social ties. Changes and transitions in individual behavior, including family and work life, could provoke the acceptance of extremist beliefs and behaviors. This is consistent with life course theories that outline different trajectories of involvement in criminal activity over time (Sampson & Laub, 1995, 2005) and general strain theory that argues that the loss of positively valued stimuli could lead to various adaptations, including criminal behavior (Agnew, 1992, 2010). This includes the potential adoption of extremist beliefs to deal with the strain caused by these life changes. An example includes Blee’s (2003) research on far-right female extremists, which found that many experienced life crises similar to “unfreezing” before they joined the movement. Similarly, Sampson and Laub (1995, 2005) identified turning points such as getting involved in a stable relationship and marrying, achieving educational success, and obtaining stable employment. If these or other key events do not occur, individuals become less connected to society and its institutions, increasing the likelihood of extremist beliefs and behaviors. 42 This section focused primarily on the development and adoption of individual extremist beliefs as a precursor to engaging in criminal behaviors resulting from those beliefs. Specific to this study, the goal is to understand why some individuals who hold extremist beliefs as either a precursor to or that result from their engagement in financial crime schemes. Also important is why various individuals that do not hold extremist beliefs choose to engage in financial crimes with those who do. For some, beliefs can be more general and occur on an individual basis, while at other times they occur within the context of extremist groups, where social cohesion amplifies these beliefs and increases the likelihood of individual and collective action, including both financial and violent crime. The key is the shift from belief, which is more likely and held by a greater number of people, to action, which is less common than belief but central to extremist crime (McCauley & Moskalenko, 2014). The development of increasingly radical beliefs and behavior occur through two separate processes and are not necessarily linked together, but the alignment of the two substantially increases the likelihood of extremist crime (McCauley & Moskalenko, 2014), including financial crimes. The next section shifts the focus from the development of extremist beliefs to focusing on criminal actions based on those beliefs. From Belief to Behavior Behavior stemming from extremist belief can be considered in the context of broader participation in far-right social movements. Social movements can be defined as “network[s] of informal interactions between a plurality of individuals, groups and or organizations, engaged in a political or cultural conflict, on the basis of a shared collective identity” (Diani, 1992, p. 13). They generally consist of a series of loosely associated events across time and space with some thread tying them together. The general movement and the organizations that share their 43 common ideology promote extremist beliefs but do not exhibit control over how individuals who adopt these ideologies will ultimately behave (Freilich, Pichardo Almanzar, & Rivera, 1999). The concept of a social movement encompasses the various aspects of anti-tax ideology and criminality, which stems from beliefs in an illegitimate government and illegal, oppressive tax system. Beliefs and behaviors are not static but fluid, reflecting similarities borrowed from various far-right extremist philosophies as individuals build on and draw from their interactions with others in different situational contexts (Chermak, 2002). This also applies to those seemingly acting alone (lone wolves), who are likely to have some loose association to the broader movement, where ideas driving anti-tax ideology (including the wide variety of common frivolous anti-tax arguments) are developed, engrained, and sustained (Moskalenko & McCauley, 2011). Those falling into the anti-tax movement do not have to belong to a specific group or have direct physical connections, but hold different variations of anti-tax ideology. While many people may hold these general beliefs and are predisposed to taking action as a result, only a small portion will engage in political activity and a smaller portion in criminal activities to follow their anti-tax ideology. As (2012) noted, many who hold extreme views and join extremist groups do not commit acts of violence that would traditionally be labeled as terrorism, nor are they involved in other types of criminal activities. However, those promoting anti-tax arguments are often somewhat different, as they are encouraging others to break the law by using fake financial instruments and refusing to pay taxes. While some simply promote beliefs about the illegality of the tax system, it is easy to cross the line into illegal behavior, which could lead to government intervention. Extremist crime networks by definition revolve around illicit activities, although not all persons are involved to the same extent at the same times, with some in the network never actually violating any criminal laws. Some are more 44 involved in criminal activities than others, while other individuals or groups are more central to the main criminal activities in the movement. Extremist behavior depends largely on interactions with other movement participants in a temporal context. Klandermans, Staggenborg, and Tarrow (2002) identified three stages of movement participation: (1) generation of mobilization potential; (2) transformation of mobilization potential into actual participation (both targeting and motivation); and (3) sustained participation and withdrawal from participation. These processes differ depending on those involved and the type of movement. Participation patterns also shift over time and across situations, both at the individual and aggregate levels. Theoretical explanations for dynamic social movement participation currently lack development, although modern theories have employed increasingly sophisticated constructs and methods to examine participation in social movements (Klandermans et al., 2002). Numerous theories have been proposed to explain participation in far-right social movements. For instance, mass society theory focuses on changes in the prevailing social structure and weakening attachments to existing institutions, leading to the adoption of extremist beliefs and participation in extremist related behavior (Halebsky, 1976). Dyer (1997) found that weakening of social institutions and economic hardship in rural farming communities was blamed on the government, giving rise to the adoption of more radical views and the acceptance of extremist groups. This is consistent with social disorganization theory (Shaw & McKay, 1942), where those in disorganized areas that lack social cohesion are more likely to adopt alternative means of coping that outside of those approved by mainstream society, including joining extremist movements and engaging in criminal behavior (Freilich & Pridemore, 2005). 45 Some empirical support has been found for these explanations when applied to far-right extremists, particularly in studies of the militia movement (Freilich & Pridemore, 2005) Additional explanations focus on economic conditions. Some have emphasized economic factors, including increased inequality, breakdown of traditional rural community institutions, and threats to economic security from globalization to explain increased right-wing movement activity in the 1990s (Dyer, 1997). Others argued that individuals of lower socioeconomic status are more likely to engage in far-right extremist behaviors due to limited education, economic security, and isolation from other counterbalancing influences (Lipset & Raab, 1970). Unlike prior studies, Piazza (2015) recently found no support for economic variables in a state level examination of various factors purported to influence extremist behavior, including poverty, the decline of manufacturing, and the decline in the security and prosperity of rural farming. While economic explanations have not extended well to the various types of far-right movement activity and have received mixed empirical support, it may well be a key driving force behind the movement participation at the individual level (Freilich, 2003). Political climate is another important potential explanation for extremist movement participation. Piazza (2015) found that Democratic Party control of the U.S. Presidency was related to increased motivation for far-right extremism, but not Democratic Party control of state government, the growth in the non-white population, or the growth in average federal income tax rates. Changes impacting white male privilege, including increased abortion rates and female labor force participation, have been found to be associated with increased levels of far-right extremist behavior (Piazza, 2015). Other political and societal influences such as the gun (O'Brien & Haider-Markel, 1998), paramilitary (Freilich, 2003; Gibson, 1994), and fundamentalist religious (Dyer, 1997; Mason, 2002; Smith, 1994) cultures dominated by white 46 males (Chermak, 2002; Ezekiel, 1995, 2002; Mitchell, 2002) have been linked to the adoption of far-right extremist beliefs and behaviors. Empirical support for these political and societal indicators is mixed (Freilich & Pridemore, 2005). Although movement participation is indicative of distrust in government and support of extremist ideology at least to some degree, it remains unclear how this relates to changes in the broader political climate (Freilich, 2003). Motivations to engage in extremist crime behavior can also change over time and across situations. Situational opportunities do not simply encourage motivated offenders to act, but trigger temptations that encourage criminal behaviors (Wortley, 1997). At the most basic level, extremist financial crimes are committed by those with both ideological and profit motivations (Belli, 2011). A single motivation alone is insufficient to explain how these crimes occur and how the networks supporting them are sustained. Offenders have various and complex motivations and their behaviors reflect these motivations. Involvement with the anti-tax movement could trigger the development of financial scheme behaviors with an ideological drive. Alternatively, interactions with the IRS over tax disputes could also lead individuals to seek out anti-tax proponents with the goal of avoiding taxes. The ideological motivation then unfolds as the individual becomes more deeply entrenched in the anti-tax movement, which could lead to participation in other related financial schemes. It is important to consider how opportunities and motivations vary depending on the situational context. Mobilization to engage in increasingly deviant behavior occurs through social interactions in networks of like-minded individuals that confirm and promote extremist belief systems (Sageman, 2004, 2008). Of central importance in this context are theories of social selection and social influence. Social selection deals with actors who choose one another to interact with based on certain characteristics (McPherson, Smith-Lovi, & Cook, 2001; Robins, 47 Elliott, & Pattison, 2001), while social influence explains how actors influence one other’s beliefs, attitudes and behaviors (Friedkin, 2006). Both of these processes are related to homophily, the idea that similar people will gravitate toward one another and become increasingly similar over time through sustained interactions (McPherson et al., 2001). Byrne (1971, 1997) argued that relationships are based on these similarities, with attraction being greater between similar others. Homophily has been used to enhance the understanding of cooffending. For instance, van Mastrigt and Carrington (2014) found strong co-offending patterns based on sex and age, particularly among young males. A central theoretical question is whether influence or selection is the primary driving force behind social cohesion. Several scholars have argued that both processes are not mutually exclusive, but inherently intertwined and play off one another (Erickson, 1988; Kandel, 1978; Leenders, 1997). These intersections of influence and selection are important for understanding how people come together in cohesive groups to engage in crime. To carry out extremist financial crime behaviors, individual offenders take advantage of the various opportunities developed through relationships with potential co-offenders who tend to be similar to them. Criminologists have long recognized relational elements of criminal behavior, where individuals engage with others to commit crime (McGloin & Nguyen, 2014; Reiss, 1986). Profit-oriented criminals look for opportunities in various industries, either to gain a stable economic foothold to use as a front organization or to funnel money through a business with the goal of making crime both predictable and profitable (Block, 1991). Others motivated by political or religious ideology require similar opportunity structures to support their activities, including committing more traditional crimes with others, regardless of whether or not they hold 48 similar ideological views. These relationships are of course limited by temporal, spatial, and structural constraints. Network theories construct the social world in terms of relational networks as opposed to a priori groups, as groups are formed through local network interactions (Wellman, 1988). The appearance of organized structures is not a pre-planned and determined phenomenon, but emerge through dynamic processes of interactions among individual actors (Morselli, 2009). Explanations of behavioral structure and composition is not assumed but developed by theorizing about the implications of relationships between social actors. Those who interact more consistently with one another while not interacting with others form the groundwork for the development of a cohesive subgroup (Frank, 1995). This idea of group influence on individual belief and behavior has long been central to many sociological and social psychological theories. Once groups are formed, there are benefits to remaining part of the group as social norms have been established that encourage solidarity among members (Frank & Yasumoto, 1998). This is particularly true when the group is faced with external threats or isolation from other members of society, which increases the power of the group to influence belief and behavior through pressures for conformity (McCauley & Moskalenko, 2008, 2011). It is logical to expect those who have already shown a commitment to their inter-group to remain cohesive over time. Only longitudinal examinations of theses processes can determine the dominant forces at work in forming and sustaining cohesive subgroups. Some subgroup members form connections with other subgroups (and those unconnected to subgroups) to attain benefits such as additional information or resources, thus increasing their own power (Burt, 1992). This benefit is unavailable within a cohesive subgroup, where information and resources tend to become redundant and existing beliefs/behaviors reinforced. 49 Connections to weaker ties outside of the context of a subgroup allows for an increased capacity for action based on access to additional resources and information (Granovetter, 1973). These connections could include individuals with unique knowledge of criminal techniques or strategies to execute different types of financial schemes. Subgroup members may need someone to fill a specific role in their scheme, such as an accountant or legal expert. Several group studies of criminal network have supported these claims. For example, Iwanski and Frank (2014) found that co-offenders would first establish groups based on common crime types, and then make connections with other groups over time to fulfill numerous additional goals. By establishing these bridging ties, additional resources are obtained while group cohesion is maintained. Thus, maintaining subgroups and establishing external links substantially enhances the capacity to undergo numerous types of financial crimes with varying levels of complexity. Another important theoretical consideration regarding relational links between actors, both generally and in the context of subgroups, is network dynamics. Networks connections unfold over time and do not persist permanently but fluctuate over time. In establishing the groundwork for an increased focus on network dynamics, Emirbayer (1997) argued that social actors respond to situational opportunities and constraints, as individual relationships are embedded in time and space within a social context that facilitates interactions. In this network perspective, relational connections unfold in a dynamic fashion based on specific situational elements (Emirbayer, 1997). The resulting social structure is not a static determinant of human behavior, but rather the constant process of interaction and negotiation results in the development of social structure, which then influences behavior and thus further alters the structure. The two influence one another in a reciprocal relationship. 50 Conclusion This chapter examined theories explaining different aspects of how individuals come to embrace far-right extremist beliefs and become engaged in far-right extremist financial crime. The next chapter will utilize key principles from these theories to outline the methodology for this network study of far-right anti-tax extremists involved in financial schemes. 51 CHAPTER 4: METHODS This chapter outlines the mixed method, multi-level, longitudinal network analysis of financial schemes linked to American far-right anti-tax movement. Goals of the study are to identify the composition, organization, and motivation of financial schemes and criminal offenders and how their networks change and evolve over time. The chapter includes sections outlining the research design, data collection techniques, operationalization of key variables, and analytical approaches. Research Design This study uses the data from the EFCDB, which comprises all financial crimes committed by political and religious extremists in the United States from 1990 to the present (Freilich et al., 2014). The EFCDB includes rich qualitative open source case histories of crimes involving many extremist and non-extremist suspects coded in a relational database capturing both quantitative and qualitative characteristics of financial schemes and criminal suspects. This study takes advantage of these unique qualities by combining the strengths of these approaches with social network analysis to examine the characteristics of how extremist tax protesters organize to engage in financial crimes. First, the inclusion criteria, data collection, and coding procedures of EFCDB are described, along with the sampling criteria for the current study. Second, the two units of analysis are defined: financial scheme and criminal suspect. Third, the analytical approach is outlined, including the application of network analysis to EFCDB data, including searching for cohesive subgroups and examining relational connections over time. Data The current study stems from the broader United States Extremist Crime Database (ECDB) study. The ECDB is an open source, relational database consisting of violent incidents, 52 financial schemes, and extremist groups. The violent incidents include homicides committed by extremists; bombings and arsons committed by environmental and animal rights extremists; and violent plots inspired by al Qai’da and associated extremist movements (Freilich et al., 2014), while the group database includes domestic right-wing extremist groups (Chermak et al., 2013). Initial data collection, database design, and variable creation occurred through the identification of relevant cases by Principal Investigators Joshua Freilich, Steven Chermak, and then Project Manager Roberta Belli. The present author assumed managerial responsibilities for the Financial Crimes project (EFCDB) two years after initial development. The first prong of the ECDB inclusion criteria is attitudinal, requiring that, at the time of the crime, one of the perpetrators subscribed to elements of a far-right extremist belief system. ECDB defines the domestic far-right as having the following ideals (Freilich et al., 2014): • “fiercely nationalist (as opposed to universal and international in orientation); • anti-global; • suspicious of centralized federal authority; • reverent of individual liberty (especially right to own guns, be free of taxes); • belief in conspiracy theories that involve a grave threat to national sovereignty and/or personal liberty; • belief that one’s personal and/or national ‘way of life’ is under attack and is either already lost or that the threat is imminent (sometimes such beliefs are amorphous and vague, but for some the threat is from a specific ethnic, racial, or religious group); • belief in the need to be prepared for an attack either by participating in or supporting the need for paramilitary preparations and training or survivalism.” 53 It is important to note that the mainstream conservative movement and Christian right are not included in this definition. The second prong of the ECDB inclusion criteria is behavioral and requires perpetrators have been indicted for criminal activity occurring in the United States. The charges must have been issued in a United States court of any jurisdiction between 1990 and 2010. Generally, cases where the prosecutor drops charges are included in ECDB, but charges dropped by the police are not included. The charges must be related to a financial scheme, such as tax fraud, money laundering, investment fraud, check fraud, or false financial liens (more information on identification and coding of financial schemes is outlined below). The EFCDB contains financial crimes committed by far-right extremists.6 Far-rightists have committed over 609 financial schemes in the United States involving over 1345 individual perpetrators, 72% (n=969) of which are extremists and 20% (n=264) non-extremist collaborators. Adequate information on the remaining 8% (n=112) is unavailable. These schemes consist primarily of tax fraud, false liens, and banking (check and loan fraud) crimes (Sullivan et al., forthcoming 2015b). Sample Selection This study focuses on a subset of far-right extremists captured by the EFCDB: anti-tax extremists (also referred to as tax protesters). While all tax protesters are considered far-right extremists for the purposes of this study, not all far-rightists are necessary tax protesters. Specifically, tax protesters believe in some of the following frivolous legal arguments (outlined in greater detail in chapter 2): 6 EFCDB also collected data on financial and material support schemes committed by jihadi extremists. Schemes associated or sympathizing with al Qaeda and affiliated movements (AQAM) committed 150 financial and material support schemes from 1990 to June 2014. These schemes were involved 286 perpetrators and consisted mostly of ideologically motivated monetary and material support for terrorism (Sullivan, Freilich, et al., 2014). 54 • income tax is unconstitutional; • labor is property and wages are not taxable; • Sixteen Amendment to the Constitution was never ratified; • tax laws don’t apply to national born “sovereign citizens”; • income tax is involuntary servitude outlawed by the Thirteen Amendment to the U.S. Constitution; • the Internal Revenue Code is not law; • income tax is voluntary. Due to being centered on tax related protest and denial, the commonalities among the beliefs held by tax protesters, and their criminal actions consisting overwhelmingly of tax avoidance and similar financial schemes, the anti-tax movement is unique to study separately from far-right extremists as a whole. The anti-tax movement consists of both extremists relying on frivolous tax arguments while being unaligned with other far-right orientations and those who are primarily associated with a different far-right orientation, such as sovereign citizens, white supremacists, or militia and patriot groups. All schemes (tax avoidance and other types) involving anti-tax extremists are included in the study in order to examine the full scope of their engagement in financial crime activities. This study utilizes an “open system” network data collection strategy, meaning every offender and scheme meeting the broadly outlined inclusion criteria is part of the study as part of the whole network, regardless of whether or not connections between components are present. In network analysis, rules of random sampling and independence of observations do not apply, as members of the network are inherently connected in some way (Wasserman & Faust, 1994). Therefore, a specific set of boundary criteria is selected and all the nodes fitting within those 55 boundaries are included in the network. In this study, those schemes and individual criminal perpetrators qualifying under the criteria outlined above are included in the network. The years 2002 through 2004 were chosen for the purposes of manageable data collection in order to sufficient time for the criminal cases to be processed through the criminal justice system and the appropriate information to be written, disseminated to the public, identified through ECDB data collection strategies, fully searched, and coded, and evaluated. Multiple years were selected to ensure an adequate initial sample similar to the overall content of the entire dataset containing all tax protesters. Sufficient data was also needed to compare the scheme and perpetrator activities across multiple time points and appropriately evaluate their emergence, evolution, and desistance. A prior study by Belli (2011) also sampled 2004 EFCDB data that included tax protesters. These data differ from Belli’s sample in several ways. First, this study focuses on active schemes during 2002 to 2004 as the initial sample, whereas Belli started with cases indicted in 2004. The cases in this sample could have been indicted anytime between 1990 and 2010; the scheme and the suspects involved are included in the study provided that the actual scheme itself was active at some point between 2002 and 2004. While this requires a more subjective judgment than indictment dates, estimates of scheme start and end dates were identifiable in almost every case. Where additional time periods for a scheme were possible but not identifiable, only the known year (typically the year of indictment) as counted as active. Second, Belli included only federal cases, while this study includes all cases (federal and state) meeting ECDB’s inclusion criteria. Third, Belli focused on all far-right extremists, while this study focuses on a subset meeting the additional criteria of being a tax protester. Fourth, while 56 Belli focused only on the individual perpetrators, this study includes both scheme (event) and suspect (individual) level data (see below for further details on units of analysis). Based on these criteria, the initial sample included 167 schemes and 331 non-unique individual perpetrators. These are non-unique because an individual involved in more than one scheme is coded separately for each scheme. In order to include the full scope of the networks involved, all additional schemes these individuals participated in were added to the initial sample along with an additional perpetrators involved in these additional schemes. This produced an additional 48 schemes and 113 non-unique perpetrators. The final study universe consisted of 215 schemes and 368 unique individual perpetrators.7 Figure 4.1 illustrates the number of schemes in the study universe (blue line) compared to all schemes in EFCDB involving tax protesters (red line) from 1990 to 2010. The graph indicates that the sample captures a substantial portion of the active schemes and roughly mirrors the basic trend lines for these years, with the most active period being between 1998 and 2003. However, given the sampling strategy and the nature of financial scheme detection, these trends should be interpreted cautiously. The study universe peaks in 2002 and declines steadily afterward. This is largely a reflection of the sampling strategy that first identified cases from 2002 to 2004. The sharp decline is likely due to a time lag in detection, investigation, and indictment of additional schemes that typically take many years to develop and are not widely known to the public until much later. It may also reflect a general downward trend in the 2000s, which is consistent with similar findings from studies examining far-right violence (see Gruenewald et al., 2013d; Spaaij, 2010). Indeed, these active scheme counts for the latter half of the 2000s will likely increase in the future as additional schemes are uncovered and reported, which could alter these trends. 7 In numerous cases, the source materials referred to a certain number of individuals involved in the scheme but did not mention their names. While these unknown suspects are coded in ECDB, they were removed from the current study if their identity could not be established. 57 Figure 4.1. Financial Schemes Involving Tax Protesters by Year: 1990 to 2010 200 180 160 140 120 100 80 60 40 20 0 All Schemes Study Universe Data Collection The following section outlines the data collection strategies and processes for the ECDB, including open source searching and coding. This includes a discussion of the two units of analysis and reliability of the data. Open Source Searching The ECDB relies on open source data collection methods to obtain information on financial schemes and assembled into a relational database (see Freilich et al., 2014). The first stage of data collection involved reviewing over sixty distinct sources, including known databases, court documents, media records, and watch-group reports, to identify financial crimes meeting the ECDB inclusion criteria (see Appendix A for a complete list of these sources). An initial list of these publically available cases was created. Each case was treated as an individual case study for the purposes of identifying relevant schemes and suspects involved in each scheme. Official records (court, government documents), existing terrorism database, watchdog 58 group reports, videos, media reports, websites, blogs, materials produced by extremists, and scholarly accounts were mined to gather all available information about each case (see Appendix B). Digital copies of the information were assembled and archived in search (word processor) files, which is essential to capture the information for subsequent use by coders and eliminate the potential for information to be lost after being removed from the search databases. The advantage of open source research is the ability to tailor the data collection to the specific needs of the researcher. While the documents are not written with the researcher in mind, the ability to collect information from a wide array of publically available sources provides a more complete picture of the issue and addresses potential biases inherent in any individual source (Lafree, Dugan, Fogg, & Scott, 2006). Open source databases are a widely used and effective way of overcoming the weaknesses of many traditional data sources (Chermak, Freilich, Parkin, & Lynch, 2012). There are several notable challenges with the use of open source information. When assessing the available information, it is important to determine the credibility of the source, particularly when gathering information from websites and posts on media sites that are open to public comment (Noble, 2004). The information uncovered in the search materials come from various sources that occasionally contains conflicting information, implicating reliability issues. These open source documents stem from the subjective views of numerous authors with their own potential bias and perspectives. In these situations greater weight was granted to the more "trusted" source. Similar to Sageman (2004, p. 65) "in decreasing degrees of reliability... [we favor] court proceedings subject to cross examination, followed by reports of court proceedings, then corroborated information from people with direct access to information provided, uncorroborated statements from people with that access, and finally statements from people who 59 had heard the information secondhand.” Appendix C lists the source types in decreasing degrees of reliability (Freilich et al., 2014). It is also important to keep track of the information found in each source over time into order to assess the quality, reliability, and consistency of the sources (Noble, 2004). Coding Coding involves reviewing the open source materials, judging whether the ECDB inclusion criteria have been satisfied, and coding the information in a relational database with multiple units of analysis. For each case study, coders first created a timeline and listed the number of schemes, suspects, and businesses meeting the ECDB criteria. If a scheme has multiple suspects, each suspect is coded. As long as one suspect meets the ECDB attitudinal criteria outlined above, every suspect participating in a financial scheme with that individual is coded. Cases not meeting the ECDB inclusion criteria were subject to further review by the EFCDB Project Manager and Principal Investigators. Once a coding decision was made, coders searched each scheme and suspect to ensure the original search found all relevant information and nothing important was missed. A relational database (Microsoft Access) was used to code the data into distinct databases: scheme (i.e. information on the criminal events and activities themselves), and suspect (i.e. known perpetrators involved in the scheme). Unique identification (ID) numbers were assigned to each scheme and those numbers were linked to each suspect involved in the scheme. This is an innovation extending beyond the majority of efforts focusing only on a single unit of analysis and allows for analyses across units. This study utilizes two units of analysis: (1) financial scheme and (2) criminal suspect. 60 Financial Scheme EFCDB defines a financial scheme as an illicit financial operation involving a set of activities (i.e. techniques) carried out by one or more perpetrators to obtain unlawful gain or other economic advantage through the use of deliberate deception. This can consist of any number of violations of the criminal code, such as failure to file a tax return, mail/wire fraud, or interfering with the administration of internal revenue laws. The scheme is distinguishable from a crime and technique, which are coded as separate variables. The crime is the actual criminal charges issued against the perpetrators. The technique consists of the specific actions taken by the perpetrators in carrying out the scheme. The scheme is the overall operation with a set number of perpetrators and objective over a specific period of time and identifiable location(s). For example, six tax protesters sold more than 150 “common law” trusts in a tax avoidance scheme called the American Asset Protection from 1993 to 1999 in Palm Beach and Martin counties in Florida, resulting in more than $2 million in lost tax revenue. This is coded as a single discrete scheme with both scheme and suspect attributes captured in the EFCDB. Perpetrators have their own distinct roles in carrying out a scheme and engage in a number of activities, or techniques, such as fundraising, laundering, creating false legal documents, administration, and marketing. Criminal charges could be numerous, including tax evasion, failure to file an income tax return, and conspiracy to impede the administration of internal revenue laws. Schemes involving different crimes, techniques, goals, suspects, time periods and locations are typically coded as distinct schemes. As noted above, the EFCDB includes 609 financial schemes involving at least one farright extremist committed in the United States from 1990 to 2013. These schemes have resulted in losses to diverse sets of victims, including citizens, businesses, financial markets, and 61 government institutions. Scheme ideological motivation deals with the totality of the entire scheme. Some are purely ideologically motivated, while others are profit-oriented. Still others have a mix of both ideology and profit/greed. If the scheme is conducted by members of an ideological group or the motivation for the scheme was to promote an ideological cause or goal, evidence for this is found in the statements of purpose made by the perpetrators or those discussing the case (i.e. police, prosecutors, attorneys, reporters, etc.). Scheme ideology may be very similar to the ideology of individual perpetrators, but these are not always one in the same. An extremist suspect may be involved in a scheme with a number of other individuals with no extremist link, making the scheme profit motivated as opposed ideologically motivated. An extremist suspect may also be involved in a purely profit-oriented scheme, such as when a white supremacist tries to fake checks at a retailer for the sole purpose of obtaining money to purchase luxury items for personal use. Determining factors in identifying the ideological connection and personal motive of individual suspects are outlined below. Criminal Suspect The second unit of analysis is the individual criminal suspect. A suspect is an individual indicted for a financial crime fitting the inclusion criteria outlined above. This is an individual who was either far-right extremist at the time of the scheme or was involved in a financial scheme where at least one of the other suspects was a far-right extremist. All suspects involved in each scheme are coded in separate codebooks to capture different characteristics of the individual criminal suspects. Both ideological extremists and non-extremist collaborators are included in the study as long as the individual is indicted for financial crime activities. 62 Ideological and Greed Association The process of determining who can be defined as a far-right extremist warrants further attention. Numerous indicators of an extremist connection, or “pro” evidence, are identified and documented from the open sources to demonstrate the association with an extremist ideology. To do this, specific information about beliefs are mined from the open source documents determine how the suspect fits the description of a far-right extremist. Specifically, personal statements, literature, music, or tattoos (e.g. swastika) can all serve as evidence of an extremist belief system. Actions are more difficult, as we try to avoid inferring beliefs based on actions. Membership in an extremist group is considered evidence of an extremist belief system. However, actions taken as part of the scheme are viewed with greater caution. However, some tax refusal crimes, filing of false liens, and using fake financial instruments can also serve as pro indicators if the activities undertaken closely follow an established pattern of behavior typical of tax protesters or sovereign citizens. In these cases, the open source materials are closely examined to determine whether or not the suspect and scheme qualify for inclusion in ECDB. Where these suspects do qualify, “con” evidence is typically recorded to establish the weakness of the extremist connection. “Con” evidence contradicts the existence of an extremist link. Examples of con evidence include mental illness, a profit or greed motive undermining the ideological connection, or little to no evidence of a direct ideological link for the suspect. Profit/greed as a con for an individual only applies when the scheme and suspect ideology are intrinsically tied together. In many cases, the only pro evidence found is directly tied to the scheme in that the extremist beliefs and arguments are used to carry out the scheme. If an individual is using sovereign citizen or anti-tax ideas or tactics in the commission of the scheme but there is also evidence the individual is using 63 these to generate substantial personal wealth, the profit/greed undermines the extremist association and is coded as con evidence. However, in the example above of a known white supremacist committing check fraud solely for personal profit, the ideological connection of the suspect is high but the scheme motive is non-ideological. In this case, the profit motive is not coded as con evidence because engaging in a scheme for profit does not undermine the ideological link for the suspect. For many suspects, no evidence of a direct extremist link can be established despite the individual being involved in a scheme with other extremists. In these cases, the suspect is coded as non-extremist. This is important because an ideological connection is not assumed and but requires direct evidence to establish. Only in cases where pro evidence is identified in the open source documents is a suspect coded as extremist. In order to distinguish between suspects with strong extremist connections and those with relatively weak connections, the strength of association assessment was created that accounts for both pro and con evidence of having the extremist connection. This measure accounts for the intensity of the affiliation to an extremist ideology using multiple pieces of information from numerous available sources (described above). Each suspect receives a score of “0” to “4.” Appendix D illustrates how EFCDB establishes this strength score.8 This same strength of association score is also used to evaluate the ideological connection of the scheme (see above for discussion of scheme ideology and motivation). Again, if any evidence directly contradicts the ideological association, such as is the case with many anti-tax schemes, this would be included as con evidence the ideology strength scale would be decreased at least a score of “2.” The score would decrease to “1” if only a single piece of pro evidence was identified. Greed could potentially undermine the extremist connection only if an egregious profit motive appeared to be 8 See Sullivan, Chermak, Wilson and Freilich (2014) and Belli (2011) for examples of the use of the strength of ideological association scale. 64 the main reason for engaging in an anti-tax scheme where the same arguments used to promote the scheme are also used to establish the individual’s extremist ideology. If not, evidence of greed would not count as a con. Simply participating in an anti-tax scheme by avoiding taxes is not counted as con evidence. An egregious profit motive is necessary to establish con evidence according to the ideological association scale. In addition to the ideological association score, an additional measure was created to assess the extent of greed or profit motivation for both schemes and individual perpetrators. This measure is similar in structure to pro and con evidence for an extremist ideology, where evidence for greed motivation is contrasted with evidence contradicting greed motivation. Again, the ideological evidence must be intrinsically tied to the greed motivation to count as con evidence. This is often the case with anti-tax beliefs being used in the course of a financial scheme that undermine the evidence for a greed motive. In these cases, greed and ideology scores may be similar for extremists involved in anti-tax schemes in particular. An individual could be high on both the greed and ideological scales, such as would be the case with the white supremacist engaged in a profit motivated check fraud scheme described above. Their greed score is high because they are interested in personal profit, but their white supremacist views do not undermine this greed motivation for engaging in the check fraud scheme and therefore would not be counted as con evidence. The same score range as the ideological strength variable (“0” to “4”) is used. Appendix E demonstrates how the strength of greed association is determined. Again like the ideological evidence and strength of association, these scores are assessed separately for each unit of analysis (schemes and suspects). 65 Reliability Several reliability issues are important in evaluating the data, including inter-coder reliability and bias from missing cases. Issues related to source reliability are addressed above. First, inter-coder reliability needs to be addressed due to the use of multiple coders. This is particularly difficult, as ECDB is a large-scale ongoing research effort with values being constantly updated as new information becomes available. However, inter-coder reliability was addressed in a number of ways. First, coders received extensive training. New project coders first coded previously coded cases and the EFCDB Project Manager compared both sets of values for consistency. Where discrepancies were found, further training and discussion were held to determine the appropriate coding. We created a listserv of ECDB personnel to share difficult issues. The EFCDB Project Manager consistently monitored the listserv and maintained consistent communications with coders to work through inconsistencies were addressed early in the coding process. Second, coding abnormalities were continually checked across coders. Again, persistent issues across coders were identified and worked through so all coders are aware of the proper coding protocol. Third, filling in values for certain ECDB variables required little interpretation as the variables captured basic facts such as a suspect’s race, age, or gender. Fourth, a database analyst subsequently validated all existing records, verifying that coders systematically applied the coding rules when creating relational records for schemes and suspects. Inconsistences were corrected to reflect the correct coding procedures. Fifth, coding is continuously reviewed for errors and updated as new information becomes available to minimize missing data and selectivity biases (Chermak et al., 2012). Sixth, as multiple research assistants originally coded the data, the current author examined coding reliability, making revisions for quality, accuracy, and consistency prior to analysis. Each case was subsequently reassigned to a 66 second coder to evaluate specific variables and complex coding issues. The coding decisions from all the coders and the EFCDB Project Manager were evaluated collectively and the most appropriate and accurate information was recorded in the database. All of these measures helped addresses potential inter-coder reliability issues. Another potential issue exists with missing cases (except when missing completely at random) producing biased results (Parkin, 2012). Information on individuals that were not highly active or did not participate in a highly publicized criminal scheme may be limited. This is particularly problematic when dealing with financial crimes, where the “dark figure” is vast and impossible gauge with any degree of accuracy. As opposed to homicides that are more likely to come to the attention of law enforcement, be subject to thorough investigation, and reported by the media (Chermak & Gruenewald, 2006), the discovery of financial crimes is comparatively inconsistent. The can only be addressed by making every reasonable effort to find all the information available in order to triangulate the sources and increase the accuracy of the data (Noble, 2004). An issue related to missing cases is likelihood of having inaccurate portrayals of extremist ideological connections. This could happen in a number of ways. First, an extremist link could either go undiscovered or not reported in the source document. Second, an extremist connection could be portrayed inaccurately or incompletely, making the ideological link appear to be different than it actually is. Third, individuals could be portrayed as having an extremist connection when one does not exist. This could be due to assumptions made by the author of the document based on the individual’s behavior or relations with others who may hold extremist views. These potential inaccuracies do not discount the value of data gleamed from open sources like the ECDB, but rather provide context for how the data should be interpreted (Parkin, 2012). 67 Analysis Mixed method, multilevel, longitudinal network analysis is used to examine the structure and characteristics of both schemes and suspects in the far-right extremist anti-tax movement. The data were exported from EFCDB into various data formats to conduct the different analyses. This section describes these general analytical approaches in greater detail. Descriptive Analysis The study begins with a descriptive analysis of the financial schemes involving far-right extremist tax protesters. Frequencies of the different variable attributes are presented for both scheme and suspect level characteristics. This descriptive analysis will provide the context for the more in depth network analyses. All statistics were calculated using Stata (Version 12). At the scheme level, the following variables from the EFCDB are examined: • Type: A number of financial crime scheme types involve far-right extremists, including tax avoidance, false liens, and check, bank, and investment fraud. • Motive: Scheme motive refers to the overall goal or purpose for carrying out the scheme (ideological, profit, or mixed ideological and profit). While there are various other goals to schemes activities, the focus here rests on the contrast between ideology and greed. A scheme is considered primarily ideologically motivated if it is committed to further a far-right belief system or to finance the activities of a far-right extremist group. Scheme are not considered primarily ideologically motivated if they are related to a far-right extremist goal but are not primarily driven by the desire to further its extremist belief system. These are a hybrid between schemes that are ideological and non-ideological in motive, representing a middle group between these two broader categories. Other schemes involve at least one far-right extremist but are 68 driven purely by non-ideological goals, primarily greed or profit. • Length: Temporal characteristics are importance to considering the context of the scheme. It is also important for establishing the changes that occur in scheme behavior and composition over time. • Size: Size refers to the number of suspects per scheme. While some consist only of lone individuals (common in tax avoidance), others involve dozens of perpetrators. • Lone Wolf9: There are differences between those schemes carried out by a single individual (lone wolf) and those requiring coordination among multiple perpetrators. Generally, those operations requiring local adoption and coordination in networks are more complex in terms of their operation and execution. The “lone wolf” variable separates these single suspect schemes from those involving networks of multiple suspects and schemes. At the suspect level, the following variables from the EFCDB are examined: • Demographics: Demographic characteristics include sex, age at time of start and end of scheme, race, and state/region of primary residence. • Affiliation: Affiliation refers to the specific far-right extremist movement the suspect primarily associated with. While suspects often fit into more than one movement, the most prominent of these is presented here. These include: general anti-tax, sovereign citizen, patriot/militia, and white supremacist. A separate category indicates those suspects who are non-extremists. • Motive: Individual motive refers to the goals of each individual suspect involved in a financial scheme (ideological, profit, mixed ideological and profit, or another goal). 9 Although there are differences between lone wolf and loner offenders based on the attachment to an informal or formal group (Gruenewald et al., 2013d), this study does not distinguish between the two and simply identifies those who engage in schemes by themselves versus those who co-offend with others. 69 This variable is used to differentiate the goal of the overall scheme operation from the motives of individual participants. In addition, the scales outlined above were used at both units of analysis to measure the extent to which the schemes and suspects were motivated by ideology and greed/profit. As indicated above, both units of analysis are independently assessed for evidence of ideological and greed motivation, receiving a score from “0” (low) to “4” (high). Some are high on one scale and low on the other, while others can be high on both scales. Those schemes and suspects that have mixed ideology and greed motives may be fairly high in both ideology and greed, but not as high as those where the available evidence points to either motive as the main driving force behind their behavior. There may also be cases where an individual perpetrator is on the ideology scale and high on the greed scale, as their scheme activities did not in any way contradict their ideological connection. These variables are discussed at length along with the motive variables for both units of analysis in order to obtain a more complete picture of the forces driving both scheme and suspect beliefs and behaviors. Two-Mode Social Network Analysis Following the descriptive analysis is a two-mode (level) social network analysis simultaneously examining both the event (scheme) and individual (suspect) levels of analysis. While traditional social science studies focus on individual unit attributes, network analysis centers on social relations, or attributes of connections between individual units (Borgatti & Everett, 1997). The central tenant of the networking perspective is that social relations are structured in ways that impact behaviors at multiple levels of interaction (Wellman, 1988). Instead of focusing on commonalities in variable attributes, network analysis provides a number 70 of statistical methods for examining relational data linking common actors and events for structural patterns between interdependent units (Wasserman & Faust, 1994). While traditional social network analysis examines connections among a single level of analysis (individuals, organizations, communities, nations, etc.), two-mode (affiliation) networks model persons participating in common events or experiences (Breiger, 1974). Data are represented in a relational matrix with each individual data point represented as xij, where i represents the individual and j represents the event in which the individual participates. These subscripts refer to the different modes, or sets of entities being examined, which correspond with the units of analysis (Borgatti & Everett, 1997). Network data can be used to represent any number of relationships, but the relationship of interest in this study is the criminal tie, where individual suspects co-offend together in financial crime schemes. ECDB relational data are inherently structured to support a two-mode network analysis. Individual perpetrators are nested within the schemes in which they participate. By design, suspects do not interact directly with one another but through participation in a financial scheme meeting the EFCDB inclusion criteria. Therefore, two-mode network analysis is an ideal approach to use with these data. One of the major problems with examining covert networks is missing links, which can dramatically change the structure of the network and various actor and network level characteristics. However, the use of two-mode data counters this limitation by only including cooffending connections within the context of each scheme, as opposed other types of connections that are more susceptible to missing data. The network connections are between individual perpetrators and the schemes they are involved in instead of linking suspects and schemes directly. An extensive effort has been undertaken with the EFCDB to capture all the schemes and 71 suspects meeting the inclusion criteria. Therefore, it is improbable that any publically known scheme and suspect information available in the open sources has not been captured in these data. Of course, this does not change the fact that there are countless scheme and connections between schemes and suspects that are not reported and are therefore unknown to the researcher. However, the structure and comprehensiveness of the EFCDB data are ideal for performing a two-mode network analysis and a realistic starting point for examining a whole network of the anti-tax movement (as opposed to egocentric network data). The exclusive focus on co-offending relationships in the network analysis could also be viewed as a limitation due to the focus on one type of relationship, ignoring the critical influence of multiplex relations. These relationships can form the basis for criminal relationships and sustain covert networks. For example, Belli (2011) used EFCDB data to examine co-offending, kinship, and business relationships in the same study, which provided a more complete picture of a smaller group of individual offenders by incorporating those who were not necessarily targeted by law enforcement but were crucial nodes in the offender’s ego networks. However, the strength of the current analysis centers on the incorporation of multiple units of analysis on a large set of schemes and suspects actively engaging in criminal activities in given time period. The drawback is that only one type of relationship is included in the network analyses presented in this study. The qualitative analysis is designed to address this limitation and may illuminate different types of relationships that facilitated interactions between subgroup members or connections between schemes and suspects not identified through the two-mode co-offending relational structure. There is tremendous potential in adding additional layers of relationships, but this is beyond the scope of the current study and is reserved for future research. There are numerous ways of analyzing and visualizing network data. Visualization 72 techniques used in this study are covered in greater detail in the Network Visualization section below. A network sociogram is first created utilizing all the two-mode network data to establish a broad overview of all the financial crime activities in the anti-tax movement. The analysis of the data then focuses on the identification of subgroups, or clusters. These results are then incorporated into the visualization. An in-depth qualitative analysis will follow the subgroup and visual analyses to identify potential factors influencing the development, adaptation, and desistence of these subgroups to form hypotheses for future studies. Cohesive Subgroups Examining the presence or absence of cohesive subgroups in far-right anti-tax networks addresses the second research question. Cohesive subgroups are important characteristics that uniquely account for aspects of social structure, behaviors, and attributes by focusing on persistent patterns of interaction (Frank & Yasumoto, 1998). This analysis identified clusters of those engaging in frequent interactions, the underlying premise being intra-cluster similarity and inter-cluster dissimilarity. Members of the same cluster exhibit a preference for co-offending together, offering structural advantages due to similar behavioral norms and resource sharing capabilities used to facilitate criminal activities. This study utilizes Frank’s (1995, 1996) KliqueFinder algorithm to determine cohesive subgroups. The cluster assignments represent subgroups where suspects co-offend in schemes together but generally avoiding engagement in schemes outside of their cluster. Table 4.1. Association between Membership in the Same Cluster and Participation in a Financial Scheme Scheme Participation No (yij=0) Yes (yij=1) Cluster Membership Different A B Same C D 73 Frank’s (1995, 1996) KliqueFinder algorithm is based on a model that extends Skvoretz and Faust’s (1999) adaptation of the p* framework to two-mode networks as outlined by Field et al. (2006). The algorithm assigns suspects and schemes to clusters by maximizing the odds ratio (AD/BC) represented in Table 4.1 (adapted from Field et al., 2006), where participation in a given cluster (cell D) is maximized and lack of participation in other clusters (cell A) is maximized. At the same time, off-diagonals (lack of participation within the cluster in cell C and participation outside the cluster in cell B) are minimized. The resulting cluster assignment represents clusters where suspects tend to co-offend in schemes together but generally do not participate in schemes outside of their cluster. Figure 4.2. Equation for Identifying Cluster Membership in Two-Mode Network " p(x = 1) % ij log $ ' = θ 0* + θ 1*sameclusterij $#1− p(xij = 1) '& The model involving suspect i and event j is represented by the equation in Figure 4.2 (adapted from Field et al., 2006), where sameclusterij is 1 if suspect i is assigned to the same cluster as scheme j, 0 otherwise. The goal of the algorithm is to maximize θ1*. A large value of θ1* indicates that actors have an increased probability of participation in events within their cluster. In order to test the internal validity of the algorithm to separate schemes and suspects into distinct subgroups, the approach utilized by Frank (1995) and Field et al. (2006) is followed by running Monte Carlo simulations to generate sampling distributions of random networks. The simulations are used to determine whether θ1* is different from θ1base*, a naturally occurring nonzero value of θ1* when using random data. The null hypothesis is that scheme participation is independent of cluster membership. The algorithm performs worse when there is less evidence of clustering and performs better when evaluating larger networks. Further details of how the Kliquefinder algorithm operates can be found in Frank (1995) and Field et al. (2006). 74 Network Visualization After determining the presence or absence of subgroups, a visual inspection of the network illuminated different characteristics of the network structure. Social network analysis relies on graph theory, which represents patterns of social relations mathematically through points (nodes) and lines connecting those points (edges) on a graph. This addresses each of the four research questions through visual examination of network, scheme, and suspect characteristics over time. Changes in connections among network members are examined by comparing separate network images identifying active subgroup memberships and connections each year for similarities and differences in network structures. Preserving schemes as part of the social structure offers insight into specific offending patterns. UCINET’s Netdraw program is used to develop the graphical representation of the network known as a sociogram.10 In order to represent the positions of the suspect and events relative to one another, multidimensional scaling (MDS) is used to create a crystallized sociogram (Frank, 1996; Frank & Yasumoto, 1996). Each unit of analysis is represented as an individual node: schemes and suspects. Both schemes and suspects are represented in the same graph. Squares represent financial schemes (criminal events) and circles represent individual criminal suspects. Lines connecting nodes are called edges. As part of a two-mode network, suspects are linked to the schemes in which they participate to visualize and analyze the network. The purpose of this approach is to maintain both units of analysis simultaneously in the same analysis with both schemes and suspects are represented in the same graph. Individual suspects are not linked to one another directly, but through their co-offending in a financial scheme. Distances are defined as the number of common schemes within the same position divided by the total number schemes 10 See Borgatti, Everett, and Freeman (2002) and http://www.analytictech.com/ for further explanation of the features of UCINET and Netdraw software. 75 in that position, ranging from 0 for no common schemes to 1 for common participation in all schemes (Borgatti & Everett, 1997). The same procedure was used for calculating distances between suspects. The positions of subgroups are then mapped relative to one another again using MDS, with distances defined as the proportion of suspects in one subgroup that participated in schemes in the other (Field et al., 2006). Qualitative Analysis After examining the structural characteristics of the network in terms of cohesive subgroups and visually analyzing the whole network and various characteristics, the next stage of the analytical strategy was to qualitatively examine the largest subgroups and any connections between subgroups more in depth. The results of the Kliquefinder algorithm determined which subgroups received further attention. Qualitative approaches are valuable for examining the importance of networks and how networks change over time (Hollstein, 2011). This included both the development and evolution of the cohesive subgroups. Given the lack of knowledge on the emergence and adaptation of networks, qualitative methods are ideal for examining network dynamics (Hollstein, 2011). This analysis addresses the third and fourth research questions. EFCDB source documents were consulted to determine common themes surrounding the dynamics of the network relations over time. This involved a detailed evaluation of the case histories to build explanations for the resulting subgroups and inter-subgroup connections and determine variables of importance for further analysis with advanced statistical techniques. MAXQDA11 software, ideal for mixed methods analysis, was used as a tool for conducting the qualitative analysis of the open source documents. All the relevant source materials pertaining to the largest subgroups were imported into the software. The program allows for data organization, systematic text analysis, and coding of multiple thematic categories. 11 See Lewins and Silver (2007) and VERBI Software (2015) for more details on MAXQDA. 76 Connections to additional individuals (emphasizing those already coded in ECDB) as well as any scheme connections or links to larger organizations were the primary focus of the qualitative coding. Once coding categories were identified, the software retrieved the coded information for further review. These features helped organize the large number of source documents and the resulting data, increasing the rigor of the analysis. Using methods for qualitative content analysis, an open coding protocol12 was used to identify common relational patterns in the data signifying the underlying dynamics of the subgroups. Coding categories were adapted as each source was consulted, constantly comparing and refining these interpretations in order to allow complete patterns to unfold organically from the original sources. This inductive analytical process is designed to uncover and correctly categorize key themes (Charmaz & Bryant, 2011; Katz, 1983; Rapley, 2011). While it is important to note that grounded theory is not being utilized, the open source data were analyzed using the general principles of grounded theory, with conclusions being inductively drawn from a collective interpretation of the data. Once additional connections are identified and the adaptive dynamics of the subgroups further understood these connections were incorporated into the network analysis, creating sociograms of the updated networks. Connections were added to the subgroups, including both schemes and suspects identified through the qualitative analysis of the original source materials that were already included in the study universe. By combining qualitative analysis with the quantitative network approaches outlined above, more complete and richer insights into the formation, evolution, and desistence of financial crime schemes were achieved and used to inform the development of hypotheses for future studies 12 See Sullivan and Chermak (2012, 2013) for a similar analytical approach used by the current author to examine sources used to construct newspaper articles on financial crimes and product counterfeiting. 77 Conclusion This chapter focused on the methodology for a mixed method, multi-level, longitudinal network analysis of financial schemes involving tax protesters in the United States. Systematic research on this far-right extremist financial crime is virtually nonexistent, illustrating the contributions of the current study to enhance the existing knowledge of tax protesters (and farright extremism generally), financial crime schemes, and criminal networks. 78 CHAPTER 5: DESCRIPTIVE ANALYSIS This chapter outlines the results of the descriptive analysis. The descriptive analysis highlights key characteristics of the two units of analysis: financial schemes (event level) and criminal suspects (individual level). Financial Schemes The scheme-level characteristics are examined first by looking at key variables (outlined in Table 5.1), including lone wolf and multiple offenders, number of suspects per schemes by ideology (extremist versus non-extremist collaborators), length of scheme, scheme type, scheme motive, and strength of ideological and greed connection. Table 5.1. Scheme Characteristics Frequency Percentage Min - Max Mean (SD) Size Lone Wolf Multi-Suspect 215 142 73 100.00 66.05 33.95 Length (years) Total Lone Wolf Multi-Suspect 215 142 73 - Type Tax Avoidance False Lien Check Fraud Investment Banking Other 215 176 9 9 8 3 10 100.00 81.86 4.19 4.19 3.72 1.40 4.65 - - Motive Ideological Mixed Ideology/Greed Profit/Greed 212 130 71 11 100.00 61.32 33.49 5.19 - - 79 1 – 34 1 – 26 1 – 34 7.28 (5.03) 6.89 (4.88) 8.03 (5.27) Size The first interesting finding is the vast number of lone offender, or lone wolf, schemes. Nearly two thirds (66%) of the 215 financial schemes in the far-right anti-tax movement were lone wolf schemes involving only a single offender. This has important implications. Those only involved in financial schemes by themselves do not demonstrate the need to co-offend with others in order to carry out their offenses. This suggests that the majority of the schemes linked to the anti-tax movement are rather simplistic, not requiring unique skill sets held by other people to be successfully carried out. However, this does not mean that all of these lone offenders were not involved with criminal activities with others for two reasons. First, these schemes involved unindicted co-conspirators in many cases. Second, many lone offenders were involved in multiple schemes, and many of those other schemes involved co-offenders. These network connections are explored further in Chapter 6. The remaining third involved multiple perpetrators. There is tremendous range in this category. Many schemes involved two perpetrators, such as couples working jointly to avoid taxes (often by failing to file a return or filing false paperwork) or associates or business partners engaged in financial crime schemes together. The largest scheme involved 27 perpetrators. Table 5.2 shows descriptive statistics for the number of suspects per scheme for both total and multisuspects schemes. The average number of suspects per scheme was two when all schemes are considered. However, the many lone wolf schemes heavily skewed scheme size (number of perpetrators) toward 1. When the 142 lone wolf schemes are removed, the average number of suspects rises to four per scheme (M=4.18, SD=4.09), with an even greater variability than when considering all schemes. 80 Table 5.2. Number of Suspects Involved in Financial Schemes Total Schemes (N=215) Multi-Suspect Only (N=73) Min - Max Mean (SD) Min - Max Mean (SD) Suspects 1 – 27 2.08 (2.81) 2 – 27 4.18 (4.09) Extremist Suspects 1 – 11 1.57 (1.31) 1 – 11 2.69 (1.79) Non-Extremist Suspects 0 – 19 0.50 (2.00) 0 – 19 1.48 (3.23) Table 5.2 also shows the differentiation in the number of suspects per schemes comparing all schemes to those with multiple suspects while also demonstrating interesting distinctions between the number of extremist and non-extremist suspects for all schemes and those with multiple participants. When all schemes are considered, each scheme has an average of 1.57 extremist suspects and 0.50 non-extremist suspects. However, like the total number of suspects per scheme, the large number of lone wolf schemes skews these statistics. As a scheme must involve at least one far-right extremist anti-tax perpetrator according to the inclusion criteria, no lone wolf schemes contain non-extremist suspects. After these are removed from the data, a somewhat different picture emerges. The average number of extremist suspects increases to 2.69 and non-extremist suspects to 1.48. While mean number of extremists increased by one whole person when taking out the lone wolves, mean number of non-extremists tripled. The standard deviation is much greater for non-extremist suspects as well (four times the mean for all schemes, three times the mean for multi-suspect schemes). This demonstrates that the number of extremist suspects is more stable and consistent across schemes than non-extremist collaborators. Non-extremist collaborators range from zero in most cases to nineteen in one case. A small number of schemes are disproportionately perpetrated by a large number of non-extremist collaborators. Of the twenty-seven schemes containing non-extremist collaborators, only three contained more than ten non-extremist suspects, while ten had only one non-extremist. 81 Length The next variable of interest is scheme length. Due to the difficultly of assessing exact start and end dates for schemes down to the month and day, this somewhat crude measure counts scheme activity separately for each individual year. Start and end years are much easier and more reliable to determine in most cases. In the five cases where start dates could not be verified, the end year for the scheme was coded as the start year, with a length of one. If the scheme is active during a calendar year, this is counted as one year. The start year is then subtracted from the end year for a total number of years active as the measure of scheme length. This same method is used to determine active and inactive schemes in the longitudinal examination of cohesive subgroups in Chapter 6. Table 5.1 provides descriptive statistics for scheme length. Schemes ranged from less than one year (rounded to one) to thirty-four years with an average length of 7.28 years. As with the number of non-extremist suspects, there are minor differences between lone wolf and multisuspect schemes in terms of length. Lone wolf schemes (M=6.89, SD=4.88) are 1.14 years shorter on average than multi-suspect schemes (M=8.03, SD=5.27). Much of this difference is due to a single scheme lasting 34 years, while the next longest scheme is only 19 years. The longest lone wolf scheme is 26 years. If the longest scheme is removed, the difference in average length drops 0.36 years to 0.78 years (M=7.67, SD=4.31). Therefore, length differences between lone wolf and multi-suspect schemes are too minor to make substantive inferences. Finally, Figure 5.1 illustrates the number of active schemes per year from 1990 to 2010. The blue line represents all schemes, red line lone wolf schemes, and green line multi-suspect schemes. Scheme activities in the study universe followed a clear trajectory, increasing rapidly from lows of the early 1990s to highs in the late 1990s, peaking around 2002, then gradually 82 declining to low levels in the late 2000s roughly equal to those of the early 1990s. These patterns across the twenty-year period are fairly consistent with little variation between lone wolf and multi-suspect schemes and the study universe. Of course, this is skewed somewhat by the initial sampling choices of active schemes from 2002 to 2004, suggesting this trajectory is not entirely surprising. However, as established in Chapter 4, the active schemes in the study universe roughly mirror the trajectory of all schemes involving tax protesters in the EFCDB, meaning this finding can likely be generalized to all financial crime scheme linked to the anti-tax movement. Figure 5.1. Active Schemes by Year: 1990 to 2010 160 140 120 100 80 60 40 20 0 Total Lone Wolf Multi-Suspect Type Not surprisingly, both lone offenders and those involved in schemes with others are most likely to engage in tax avoidance (82%). Percentages are consistent across lone wolf and multisuspect schemes. The majority of these schemes involved failing to file income tax returns, submitting false tax returns, and sending false documentation to officials in an attempt to obstruct or impede the collection of taxes. Other schemes involved the sale of anti-tax packages, 83 where those looking to save money on taxes were charged extravagant fees and provided fake documentation, instructions on how to file false tax documents, and fake financial instruments to satisfy existing debts. Often tax avoidance schemes involved tax protesters who decided to stop participating in the federal tax system, responding by stopping their tax withholdings and refusing to file tax documents. Due to their nature, most of these tax avoidance schemes involve various “frivolous arguments” tax protesters typically utilize as part of their ideology to carry out the schemes. The prevalence of both lone wolf and multi-offender schemes for tax avoidance speaks to wide variability in the composition and content of tax avoidance. Many involve individual offenders in their own operations, most likely individual tax avoidance. However, others are involved in more complex operations, such as selling tax avoidance products based on anti-tax and sovereign citizen arguments, which can involve a large number of individuals with different skills and roles to carry out. Again, many of these schemes are also linked together through larger common schemes of all types. Table 5.3. Financial Scheme Type: Lone Wolf vs. Multi-Suspect Tax Avoidance False Lien Check Fraud Investment Banking Other Total Total Frequency Percentage 176 81.86 9 4.19 9 4.19 8 3.72 3 1.40 10 4.65 215 100.00 Lone-Wolf Frequency Percentage 114 80.28 8 5.63 7 4.93 4 2.82 2 1.41 7 4.93 142 100.00 Multi-Suspect Frequency Percentage 62 84.93 1 1.37 2 2.74 4 5.48 1 1.37 3 4.11 73 100.00 Several other types of schemes were found in addition to tax avoidance, including false liens, check fraud, banking, investment, and several other types of schemes. Nine additional schemes (4%) involve false liens. The filing of false liens and other legal documents against public officials or other citizens (also known as “paper terrorism”) is often done in retaliation for criminal or civil proceedings brought against them or other movement affiliates. These liens typically claim billions of dollars in liabilities and are intended to both financially harm and 84 intimidate the victim. All but one of these schemes involved lone offenders. Similarly, check fraud schemes were also more likely to involve single perpetrators. Check fraud (4%) schemes typically consist of fictitious, often self-manufactured checks, money orders, or other similar financial instruments, including those utilized by proponents of the “redemption” or “strawman” theory, such as a “sight draft” or “bill of exchange.” Two check fraud schemes involved more than one perpetrator. Perhaps the prevalence of individual offenders speaks to the simplicity of false lien and check fraud schemes. These types of schemes do not require any specialized knowledge other than the anti-tax and sovereign citizen arguments that support them, meaning that anyone with access to basic information can use a fake check or file a false lien. This knowledge is easily obtained online or through anti-tax seminars commonly held across the U.S. on a regular basis. The remaining schemes had much greater variability in size, composition, and purpose. Investment schemes (4%) include pyramid and Ponzi schemes, as well as other illegal securities instruments or investment scams. Many of these revolve around debt elimination, where the perpetrators sell access to (nonexistent) secret government accounts owed where money can be used to satisfy debts. Other variations of the debt elimination scheme include selling access to the proceeds of phony lawsuits against the government, such as one where proponents declared themselves to be executors of a lawsuit settlement declaring the federal reserve bankrupt, with those paying a filing fee gaining access to millions in funds. These schemes were equally likely to have single or multiple perpetrators, but the multi-suspect schemes ranged from two to thirteen in size, mirroring tax avoidance in terms of size variability. Other banking related schemes (1%) included mortgage and credit card fraud. One of these was a mortgage fraud scheme involving five perpetrators, only two of which were 85 identified as extremists. Unlike tax avoidance, false liens, and check fraud, banking and investment schemes were evenly split between those involving single and multiple perpetrators. These speaks perhaps to the wide variety of activities involved in carrying out these more diverse and intricate schemes, but may also simply be an artifice of the low numbers of each of these schemes in the study universe. Finally, other types of schemes (5%) included filing false legal or tax documents against officials (similar strategy as false liens), embezzlement, and money laundering (including laundering of proceeds from other illegal financial schemes). Motive Table 5.4 highlights scheme motivations. The primary motivation for each scheme varied, but the majority (n=130, 61%) were primarily carried out based on a political or religious extremist ideological belief system. This represents an overwhelming number of ideological scheme motivations compared to greed motivations. Only 5% (n=11) were not motivated by ideology, but conducted instead based on greed or profit motive. The remaining 33% (n=71) consisted of both ideology and profit as motivating factors. Table 5.4. Financial Scheme Motive: Lone Wolf vs. Multi-Suspect Ideological Mixed Profit/Greed Total Total Frequency Percentage 130 61.32 71 33.49 11 5.19 212 100.00 Lone-Wolf Frequency Percentage 93 66.43 38 27.14 9 6.43 140 100.00 Multi-Suspect Frequency Percentage 37 51.39 33 45.83 2 2.78 72 100.00 These findings are further illustrated by the strength of ideology and greed association scores in Table 5.5 with high mean ideology strength (M=2.77) and low mean greed strength (M=0.55). These are due to the large number of entirely ideologically motivated schemes (ideology value of “4”) with no evidence contradicting the ideological association (44%) and the even larger number of schemes without any evidence for greed association as a contributing factor (greed score of “0”). On the reverse ends of the scale, greed strength indicates only two 86 schemes with multiple pieces of evidence for greed and no contradictory evidence (greed score of “4”). The remaining greed motivated schemes have strength scores of 3. Meanwhile, all eleven profit motivated schemes have no ideological evidence (ideology score of “0”). The extreme ends of the scale are logical according to the design, as profit schemes do not have evidence for ideology and vice versa. Any evidence to the contrary automatically reduces the strength score to “2”, resulting in a scheme with mixed motivations (mixed motivations could also be a “1” score). Table 5.5. Scheme Strength of Ideology and Greed Association Ideology Strength Greed Strength Frequency Percentage Frequency Percentage Min-Max 0–4 0–4 Mean (SD) 2.77 (1.29) 0.55 (0.87) Value 0 1 2 3 4 11 32 45 32 93 5.16 15.02 21.13 15.02 43.66 134 52 15 9 2 63.21 24.53 7.08 4.25 0.94 Total 213 100.00 212 100.00 The majority of greed scores reflect the finding of a single piece of pro evidence favoring a greed motivation. This results in scores of either “3” or “1” (“1” in cases where ideological evidence contradicts the greed motive). This is due largely to two factors. The first is the difficulty in establishing multiple pieces of greed evidence compared to the ideological, which is more clearly defined than greed evidence. The second is the comparative amount of time spent addressing ideological association compared to greed. The ECDB logically focuses vast resources on identifying extremist connections, which explains in part the large number of extremist motivations. The results of this effort are that ideological associations and motivations are more clearly defined than others. 87 Table 5.4 also distinguishes between motivations for lone-wolf and multi-suspect schemes. The distributions are faily consistent with a few notable exceptions. Namely, multisuspect schemes are more likely to have mixed motives of both ideology and greed and less likely to ideology or greed alone. Only two (3% of) multi-suspect schemes were driven primarily by greed compared to nine (6% of) lone-wolf schemes. This is a logical finding given the varying contributing influences to a multi-suspect operation compared to a lone wolf. Multisuspect schemes have a number of different individuals involved which conplicates the overall goal or purpose of the scheme, making it more likely to find evidence of both types of motivation. They are less likely to be purely ideological or greed motivated. Table 5.6. Scheme Strength of Ideology Association: Lone-Wolf vs. Multi-Suspect Lone-Wolf Multi-Suspect Frequency Percentage Frequency Percentage Min-Max 0–4 0–4 Mean (SD) 2.81 (1.30) 2.70 (1.28) Value 0 1 2 3 4 9 19 23 28 61 5.16 15.02 21.13 15.02 43.66 2 13 22 4 32 2.74 17.81 30.14 5.48 43.84 Total 140 100.00 73 100.00 Tables 5.6 and 5.7 further demonstrates these differences by distinguishing between lonewolf and multi-suspect schemes for scheme strength of ideological and greed associations, respectively. While there are generally similar distributions between lone-wolf and multi-suspect schemes, there are far greater scores of “2” for both ideology and greed for multi-suspect schemes compared to lone-wolf schemes. This further demonstrates their ambiguous nature, with evidence for both ideology and greed. Mean greed scores are similar, although mean score for multi-suspect schemes (M=0.69) is slightly higher than the lone-wolf mean score (M=0.48), a 88 further reflection of the greater number of “2” strength scores. Finally, there are fewer “3” scores for multi-suspect schemes, which may be a reflection of the greater amount of potential evidence in the source materials for schemes involving multiple perpetrators. Table 5.7. Scheme Strength of Greed Association: Lone-Wolf vs. Multi-Suspect Lone-Wolf Multi-Suspect Frequency Percentage Frequency Percentage Min-Max 0–4 0–4 Mean (SD) 0.48 (0.86) 0.69 (0.88) Value 0 1 2 3 4 96 31 4 8 1 68.57 22.14 2.86 5.71 0.71 38 21 11 1 1 52.78 29.17 15.28 1.39 1.39 Total 140 100.00 72 100.00 There are also numerous similarities across types and when comparing lone-wolf and multi-suspect schemes to total schemes. There are few differences between ideology mean scores between lone-wolf (M=2.81) and multi-suspect (M=2.70) schemes and similar proportions of “4” strength scores with no evidence contradicting the ideological association of the scheme and “1” scores with only a single piece of pro ideology evidence along with con evidence. Similarly, there are few high strength greed scores for either lone-wolf or multi-suspect, along the majority of higher scores with no evidence contradicting the greed motive are lone-wolf schemes. Finally, there are similar frequencies of schemes with no greed association (“0”) or single greed indicator contradicted by an ideological indicator (“1”). However, a greater proportion of lone-wolf schemes have “0” scores while a greater proportion of multi-suspect schemes have “1” scores. Further nuances are found when examining individual suspect motives. The next section first focuses on demographic statistics before focusing more attention on individual motivations. 89 Criminal Suspects There are a total of 368 unique individual criminal suspects involved in the 215 financial schemes discussed in detail above. Of the 368 suspects, 279 (76%) are identifiable far-right extremists while 89 (24%) are non-extremist collaborators. Characteristics of both extremists and non-extremists to be discussed in depth include demographics (gender, race, age, and geographic region), affiliation, and motivation (including strength of ideological and greed association) outlined in Table 5.8. Table 5.8. Suspect Characteristics Total (N=368) Frequency Percentage Extremist (n=279) Frequency Percentage Non-Extremist (n=89) Frequency Percentage Sex Male Female 368 287 81 100.00 77.99 22.01 279 223 56 100.00 79.93 20.07 89 64 25 100.00 71.91 28.09 Race White Black Other 327 309 13 5 100.00 94.50 3.98 1.53 259 244 11 4 100.00 94.21 4.25 1.54 68 65 2 1 100.00 95.59 2.94 1.47 Age (at start) Min - Max Mean (SD) 437 19 - 72 45.34 (9.95) - 274 19 - 68 45.94 (9.67) - 88 20 - 61 40.85 (10.14) - Age (at end) Min - Max Mean (SD) 440 22 - 78 51.59 (9.77) - 276 22 - 76 52.48 (9.60) - 88 28 - 67 47.83 (10.10) - 100.00 44.04 23.47 18.77 10.47 3.25 89 30 18 29 12 0 Region South West Northeast Midwest Other 366 152 83 81 41 9 100.00 41.53 22.68 22.13 11.20 2.46 277 122 65 52 29 9 Affiliation Anti-Tax Sovereign Citizen Patriot/Militia White Supremacist Non-Extremist 368 140 126 12 1 89 100.00 38.04 34.24 3.26 0.27 24.18 - 90 - - 100.00 33.71 20.22 32.58 13.48 0.00 - Demographics Those involved in these financial schemes in the anti-tax movement are generally middle aged, white, male individuals who reside disproportionately in the Southern region of the United States. However, there are several notable distinctions worth noting for each of these demographic characteristics. First, while over two-thirds of the suspects are male (n=287, 78%), there are a substantial number of female perpetrators (n=81, 22%) involved in these schemes. While these proportions are similar across extremist and non-extremist suspects, a notably greater proportion of females are non-extremist compared to males. While only 20% of males are non-extremist, 28% of females do not have an identifiable extremist connection. This is reflected in the numerous female perpetrators who went along with their husband’s wishes with the scheme but did not exhibit any characteristics of extremist ideology themselves. Second, race is overwhelmingly white at 95% (n=309) of all suspects. This is consistent across extremists and non-extremists. Only 18 individuals were not white, with the next most prominent race being black at 13 suspects (4%). However, race could not be determined for 11% (n=41) of the suspects. Race was similar for extremists and non-extremists. Third, suspect age had a tremendous range. Measures for age at both the start and end of schemes were recorded for each individual. Age was based on the total number of non-unique individual suspects. Each suspect was coded once per scheme in order to obtain an accurate age count for each individual per scheme, which is why the number of suspects is higher than the total number of unique individuals involved. Start and end ages could not be determined for eight and five non-unique individuals, respectively. The youngest suspects ranged between 19 (at start) and 22 (at end) years old, while the oldest were between 72 (at start) and 78 (at end). The mean 91 starting age was 45 and mean end age was 52. The mean ages for extremist suspects were nearly identical to those for all suspects. However, while mean age for extremists and non-extremists is similar, non-extremists on average were roughly five years younger than extremists. Figure 5.2. State of Suspect Primary Residence 60 50 40 30 20 Other Nebraska New Mexico New York Montana Idaho Virginia Wisconsin Tennessee S. Carolina Michigan Nevada N. Hampshire Louisiana Hawaii Utah Minnesota Massachusetts Oregon Pennsylvania Texas Colorado Arizona Illinois Washington Ohio California 0 Florida 10 Finally, suspects differed by their primary geographic location. Given the primary focus of the current study on temporal as opposed to spatial characteristics, a full accounting of the geographical locations of all schemes and suspects is out of scope. However, the primary state of residence for each suspect provides important context for where these individuals were located and where their financial crime activities primarily occurred. These are presented in two variables: state and region. Region is presented in Table 5.8 while all states with greater than three suspects as its primary residence are illustrated in Figure 5.2. Suspects disproportionately resided in the Southern region of the United States (42%). By far, the largest number of suspects resided in Florida (n=52, 14%), accounting for a large portion of the Southern region suspects. 92 Those in the Southern region are also more likely to be extremist than those in the Northeast and Midwest regions (44% of all extremists compared with 34% of all non-extremists). The Midwest region had the least suspects (n=41, 11%). The Northeast region consisted of the largest number of non-extremist collaborators (33% compared to only 19% of extremists). Lastly, several other states of note from Figure 5.2 are the three next highest ranked states after Florida: California (n=32), Ohio (n=24), and Washington (n=20). Virtually every state was represented with 42 states and the District of Columbia serving as the primary residence for at least one individual suspect involved in financial schemes linked to the anti-tax movement. Affiliation Again, of the 368 suspects, 279 (76%) were identifiable far-right extremists while 89 (24%) were non-extremist collaborators. The extremists involved in these schemes belonged to a number of movement affiliations, or subcategories of the broader far-right extremist movement. While extremists often potentially fit into multiple affiliation categories depending on the situational and temporal context (Chermak, 2002), only the most prevalent affiliation is reported here. The most prominent affiliation was with anti-tax protester ideology (n=140, 38%) with no other identifiable affiliation. These were individuals who subscribed to anti-tax beliefs, tactics, and methods but who do not qualify as sovereign citizens, patriot/militia, or white-supremacists. Sovereign citizens have nearly the same number of suspects (n=126, 34%) as anti-tax protesters. The major difference between the two is the reliance on arguments or tactics expressing or relating sovereignty from the government or citizenship in a state republic as opposed to the United States. Any suspect with this sovereign connection was coded as a sovereign citizen as opposed to general anti-tax. An additional 3% (n=12) were affiliated with militia or Christian Patriot groups, which were coded as a single category. Only one suspect was identified as a 93 white supremacist (KKK, neo-Nazi, Christian Identity, etc.), although this individual also held elements of sovereign citizen and Christian Patriot philosophies. The large number of non-extremist collaborators involved in financial schemes with extremists is a notable finding. These individuals had no identifiable connection to an extremist ideology. Another key finding is the involvement of non-extremists in financial schemes involving far-right extremists. One-fifth of the individuals charged with participating in these schemes did not have any identifiable expression of extremist ideology. Non-extremist collaborators engaged in these schemes primarily for financial gain. This includes individuals with specialized expertise providing professional services, such as tax preparers, lawyers, accountants, financial planners, and other businesspersons. They filled a number of roles including preparing and filing tax returns and preparing false financial instruments and legal documents. A few were even bank executives. In other cases, individuals joined the scheme as a financial opportunity after being introduced by family members, friends, or associates. Many involved spouses going along with their partner’s wishes, but not exhibiting any identifiable extremist ideological belief system themselves. One of these cases involved a federal government employee who claimed to be unaware that her husband was falsifying their tax returns. Finally, many non-extremists were involved with schemes or organizations with an extremist link, most particularly those using anti-tax arguments. These individuals were often marketers, salespersons, or office workers employed by the organization’s leaders, or they filled a specialized role (discussed above). Others were involved in the operations of the scheme, but there was not evidence directly tying the individual to the extremist arguments used to carry out the scheme, so they were coded as non-extremists. Typically (but not always), these individuals 94 were coded as having a greed motivation. It is important to note that many of these individuals could also have far-right extremist attitudes and beliefs, but this could not be reliably determined with the information available. The differences in motivation based on ideology and greed (including strength scales for both) are discussed further below. Motive Table 5.9 outlines the tremendous variation in motives for individual suspects participating in financial crime schemes.13 While half (n=225) are ideological, nearly one fourth (n=107) are motivated by greed. The remaining one-fifth has a mix of ideology and greed. According to how extremists are operationalized, all those with an ideological or mixed motive are coded as extremists, as no non-extremist can have either of these motivations by definition. Naturally, given the number of non-extremists, it is expected that a large portion of the greedmotivated individuals will not have an extremist association. This is indeed the case as 69% (n=74) of those with a greed motive are non-extremist. The remaining 31% (n=33) are far-right extremists who primarily engaged in financial schemes due to greed instead of their ideology. While only 9% of extremist compared to 78% of extremists motived by greed, this is still a sizeable number worth noting. In these cases, the extremist connection for the individual was established, but their motive for the specific scheme appeared to be greed, not ideology. That is, their scheme activities were not intrinsically tied to their ideology, so greed did not undermine the ideological connection and vice versa. However, the opposite is true when anti-tax arguments are used to conduct a large scale tax avoidance scheme motivated by greed. Typically, schemes involving a large number of those with greed motives are either mixed or greed motived, although they could be involved in an ideologically 13 Excludes suspects with unknown motive. As suspect motives vary across schemes, suspects involved in multiple schemes are coded separately for each unique scheme. This resulted in a total of 427 non-unique suspects with identifiable motives. 95 motivated scheme where the main perpetrators driving the scheme are ideologically motivated. Many of these extremists participated in additional schemes with ideological motives, but these others schemes were motivated by greed. Table 5.9. Suspect Motives Ideological Mixed Profit/Greed Other/Unknown Total Total Frequency Percentage 225 50.56 89 20.00 107 24.04 24 5.40 445 100.00 Extremist Frequency Percentage 225 64.29 89 25.43 33 9.43 3 0.86 350 100.00 Non-Extremist Frequency Percentage 74 77.89 21 22.11 95 100.00 The vast majority of those with other or unknown motivations (88%) are non-extremist. This is likely to due to the focus on ideological connections, where evidence for ideological motivations is more readily available and interpretable, receiving disproportional attention due to the focus of the ECDB data. As with additional scheme motivations, there are various reasons why individuals choose to participate in criminal activities. Ideology and greed are two motivations explored in this study along with network influences in Chapters 6 and 7. Nonextremists with motivations other than greed did not receive the same attention in the current study as the focus was restricted to ideological offenders and their contrast with greed motivated individuals. In addition, other motivations connected with ideology, such as psychological and social (such as group affiliation) characteristics, as well as retaliation and revenge, are usually tied with and coded as ideology unless they distinctly and notably differ. Finally, it should be noted that the three extremists with unknown or other motives were those involved in prior financial schemes in the study timeframe (1990 to 2010) where the motivation could not be established. If an individual is coded as extremist, this assignment remains consistent across all schemes regardless of motive or evidence at the particular time. If additional schemes are identified where the motivations are unclear, the motive is not assumed and coded as unknown. In almost every case, sufficient evidence existed to establish motivation. 96 Table 5.10. Suspect Strength of Ideology and Greed Association Ideology Greed Frequency Percentage Frequency Percentage Min-Max 0–4 0–4 Mean (SD) 2.37 (1.56) 1.03 (1.29) Value 0 1 2 3 4 85 49 92 35 173 19.59 11.29 21.20 8.06 39.86 223 82 46 55 26 51.62 18.98 10.65 12.73 6.02 Total 434 100.00 432 100.00 These findings are enhanced by the strength of ideology and greed association scores in Table 5.10 with moderate mean ideology (M=2.37) and greed strength scores (M=1.03) compared with the higher mean scheme ideology (M=2.77) and lower mean scheme greed motivation (M=0.55). This reflects the increased variation of individual motivations compared to scheme motive. Individuals on the whole appear less ideological and more greed motivated. This is due in large part to the number of non-extremist collaborators. The majority of schemes tend to be motivated by ideology with some involving non-extremists who go along with the scheme but do not impact its overall goal or purpose. However, schemes represent the whole operation of carrying out criminal activity with varying contributing factors to the overall goal, whereas those involved often have different motivations for engaging in the scheme. Some are motivated by greed, others by ideology. Others have entirely different motivations, such as helping out a family member or friend while claiming ignorance to the crimes being committed. For the strength of ideology, a large portion has a strong ideological connection (“4”) or mixed ideology and profit (“2”). This further provides further support for the suspect motives outlined in Table 5.9. Comparatively, few individuals have ideology scores of “3” representing only a single piece of pro evidence in favor of an ideological connection. This differs from 97 scheme ideology where a greater proportion had only single indicators, which represents the increased amount of evidence available to explain individual extremist connections, whereas this does not always translate to explaining scheme motivations. The greed ideological strength provides a contrasting view in several ways. There are more “3” and “1” scores than other scores (excluding “0” which represents extremists with no greed evidence). Those with scores of “3” are either non-extremists with only a single piece of pro greed evidence or extremists whose ideological association did not undermine the evidence of a greed motivation (not as common as the latter). The “1” score represents extremists with only one piece of evidence supporting a greed motivation and other pieces of evidence contradicting this greed motivation, namely an ideological association intrinsically tied to scheme activities. As a result, far fewer individuals have strong greed strength scores than those with strong ideology scores (6% versus 40%). Table 5.11. Suspect Strength of Ideology Association: Far-Right vs. Non-Extremist Extremist Non-Extremist Frequency Percentage Frequency Percentage Min-Max 1–4 Mean (SD) 2.95 (1.15) Value 0 1 2 3 4 0 49 92 35 173 0.00 14.04 26.36 10.03 49.57 85 - 100.00 - Total 349 100.00 85 100.00 Distinguishing between extremists and non-extremists illuminates several interesting findings. This is further reflected in the Tables 5.11 and 5.12, which distinguish between extremists and non-extremists for ideology and greed association, respectively. First, mean ideology score obviously increases when the 85 non-extremists are removed. In this case, the mean ideology score is 2.95, an increase of 0.58 from the mean score for all suspects. Again, 98 non-extremists by definition do not have an ideological association while extremists have at least one piece of evidence pointing to an ideological connection. The result is that scores for extremists range from “1” to “4” while all non-extremists automatically receive a “0” score. Similarly, greed scores for non-extremists cannot be “1” or “2” as this implies an ideological connection undermining the greed motive, which is not possible due to how non-extremists are operationalized. Non-extremists with “0” scores for greed had other motivations driving their financial crime related behavior. Second, more evidence of an ideological association is found in the increased percentage of those with a strength score of “4” while far fewer have scores of “3.” This indicates multiple pieces of evidence to support an extremist connection without any contradictory evidence were found for nearly half of the extremists. Another fourth had multiple pieces of pro evidence but contradictory evidence was found, resulting in a “2” for association strength score. Finally, only 14% (n=49) had “1” strength scores for ideology, meaning that those with an ideological connection were more likely to have multiple pieces of evidence in the source materials establishing this connection. Third, the greed strength scores are lower for extremists than the overall greed strength. The mean greed score decreased 0.46 from 1.03 to 0.57. As with the increased mean ideology score described above, this is self-evident given the removal of a large number of non-extremists with high greed scores. Interestingly though, few extremists (n=7, 2%) had high greed strength scores of “3” or “4” while 24% had “1” greed scores. This is due to the single pieces of evidence found in favor of a greed connection as discussed in the scheme strength section above, with ideology indicators more readily identifiable than greed evidence. In most cases, the ideological motivation for the suspect was closely tied to scheme activities, undermining the greed 99 association. This would not be as evident in cases where anti-tax or sovereign citizen arguments were not used to carry out the schemes. The greed strength for individual extremists involved in non-ideologically driven schemes is higher than those involving an ideological motivation. Table 5.12. Suspect Strength of Greed Association: Far-Right vs. Non-Extremist Extremist Non-Extremist Frequency Percentage Frequency Percentage Min-Max 0–4 0–4 Mean (SD) 0.57 (0.83) 2.92 (1.16) Value 0 1 2 3 4 213 82 46 4 3 61.21 23.56 13.22 1.15 0.86 10 51 23 11.90 60.71 27.38 Total 348 100.00 84 100.00 Finally, greed scores for non-extremists were much higher than those for extremists, which again is an artifact of the operationalization of the association strength variables as scores of “1” or “2” for non-extremists not possible. The overall mean score increased by 1.89 from the mean score for all suspects (M=1.03) to 2.92. Obviously, when non-extremists with other motives are dropped, the mean greed score is very high (M=3.31). However, as noted above with the large number of “1” scores for extremists, single pieces of evidence supporting the greed connection were more likely than multiple indicators. Indicators of a greed association for engaging in financial schemes will need further attention in future studies. 100 Conclusion This chapter outlined various scheme and suspect characteristics, including lone wolf and multiple offenders, number of suspects per schemes by ideology (extremist versus non-extremist collaborators), length of scheme, scheme type, suspect demographics, suspect affiliation, and motivation and strength of ideological and greed associations for both the scheme and suspect level of analysis. This descriptive overview provided the background for the study sample to further understand both the structure of the data as well as the driving forces behind the criminal behaviors. These included lone-wolf and multi-offender schemes as well as those driven by ideology, greed, or a combination of the two. This descriptive analysis set the stage for further examinations of behavioral characteristics, particularly for those co-offending in financial schemes. The next chapter focuses on network dynamics by determining the extent to which individual suspects participate in common schemes together and how these associations change over time. 101 CHAPTER 6: COHESIVE SUBGROUP ANALYSIS This chapter outlines the results of the analysis for cohesive subgroups using the Kliquefinder algorithm. The first section is a descriptive overview of the structure of the whole network. The second section illustrates the results of cohesive subgroup analysis, followed by a descriptive overview of each subgroup. Third, the subgroups are examined longitudinally over a twenty-year period to determine how they evolve over time. Finally, limitations and implications of the cohesive subgroup analysis are discussed, including implications for further research. Table 6.1. Network Components Component Size Frequency Two 69 Three 41 Four 13 Five 6 Six 4 Seven 3 Eight 2 Nine 5 Ten 1 Eleven 1 Twelve 1 Thirteen 1 Fifteen 3 Eighteen 1 Twenty-Nine 1 Total 152 Percentage 45.39 26.97 8.55 3.95 2.63 1.97 1.32 3.29 0.66 0.66 0.66 0.66 1.97 0.66 0.66 100.00 Overview of Data Structure The 215 schemes and 368 suspects in the data are separable into 152 individual network components ranging in size from 2 to 29 nodes. Table 6.1 outlines the frequencies of each size network component. Not surprisingly given the number of lone-wolf schemes, two node components make up nearly half (45%) of the network components. This supports the descriptive findings that a large number of schemes involve a lone offender disconnected from other financial scheme activity in the anti-tax movement. However, many of these lone wolf schemes 102 are attached to larger networks through the involvement of these individual suspects in other schemes, so the number of two-node components is smaller than the number of lone wolf schemes. The other components (55%) have at least three nodes, signifying financial crime activities involving more than one perpetrator or more than one financial scheme. Three node components (one scheme committed by two individuals or a single individual committing two schemes) are an additional 27% of the remaining components. Components with four or more nodes make up only 28% (n=42) of the whole network. Only one component had more than twenty nodes. Before proceeding with the discussion of the network structure and composition, it is first essential to highlight some basic features of the sociograms. Two graphs featuring all the components in the network (Figures 6.1 and 6.2) illustrate features of both ideological (Figure 6.1) and greed (Figure 6.2) motivations. Node positions in Figures 6.1 and 6.2 are determined using multi-dimensional scaling (MDS) with nodes more similar being placed closer together in the graph. Schemes are squares and suspects are circles. Various shades of red and blue are used to highlight different motivations. Red presents far-right extremist ideology while blue represents greed. This is presented differently depending on the type of node. For suspects, extremists are red and non-extremists are blue. For schemes, ideological motive is red, greed motive blue, and mixed motives purple. Suspect motives are represented through their scheme connections (edges), allowing different motives to be illustrated for involvement in each individual scheme. Like scheme motive, red lines are ideology, blue greed, and mixed purple. Suspect with other/unknown motive have light grey edges. Except where otherwise noted, these features remain constant for all graphs presented in Chapters 6 and 7. 103 Figure 6.1. Whole Network Illustrating Ideological Motivation: All Components 104 Figure 6.2. Whole Network Illustrating Greed Motivation: All Components 105 Strength of ideology and greed association are represented separately in two different sociograms. That is, each sociogram in Chapters 6 and 7 is presented twice in separate figures that highlight these different scales through the sizes of scheme nodes and edges. For graphs focusing on ideological motivation (such as Figures 6.1 and 6.3), larger scheme nodes indicate stronger scheme ideological strength and smaller nodes indicate low scheme ideological strength (scale “0” to “4”). Similarly, thicker edges represent stronger suspect ideological strength, and thinner edges indicate low suspect ideological strength (scale “0” to “4”). The same rules apply with the second sociogram only with the greed strength scale, with larger/thicker nodes/edges representing stronger greed motivation and smaller/thinner nodes/edges representing low/no greed motivation. This captures the motivational complexities outlined in detail in Chapter 5 through visual representations of the network, illustrating as much information in each graph as possible to explain the composition of the various criminal networks in the anti-tax movement and what forces are driving their underlying beliefs and behaviors. Figures 6.1 and 6.2 are limited in their analytical value but provide a “big picture” view of all activities in the whole network at one time, illustrating a few important features identified in descriptive statistics (Chapter 5). The clearest of these is the overwhelming ideological motivation for the majority of suspects and schemes. However, the large number of mixed motive schemes is apparent as well as the non-extremists. Extremists appear to be scattered randomly while non-extremists appear concentrated in a few areas near the center of the graph. This corroborates the descriptive findings with an overwhelming majority of extremist suspects engaging mostly in ideologically motivated tax avoidance. The randomness of the node placement, particularly for those who are ideological, is indicative of the two-node components entirely separate from others in the network. 106 Figure 6.3. Whole Network Illustrating Ideological Motivation: Two and Three-Node Components Removed 107 Figure 6.4. Whole Network Illustrating Greed Motivation: Two and Three-Node Components Removed 108 The graphs also begin to illustrate the wide variations in data structure as well as the large number of components of varying sizes. When two nodes (one suspect and one scheme) and three node components (one scheme and two suspects or two schemes and one suspect) are removed and the graph is rescaled in Figures 6.3 and 6.4, the network becomes clearer as several components are revealed. While the network remains fairly dispersed, non-extremists and greed motives are more prominent than with two or three node components. Non-extremists are mostly concentrated around a small number of multi-suspect schemes. Specifically, four components in the center to upper-middle of the graph involve a larger number of non-extremists involved in schemes with either greed or mixed motivations. This provides further support for the variation in characteristics of multi-suspect schemes compared to lone wolf schemes (i.e. mixed scheme and suspect motives). Namely, schemes with a larger number of individuals are more likely to have a greater proportion of non-extremist collaborators compared to extremists. This is due in large part to the unique skills needed to pull off more expansive operations but also the fact that the larger operations are inherently less cohesive. The motivations for individual involvement will vary to a greater degree than when only two or three people are involved. This chapter will now turn from focusing on the contrast between lone wolf schemes and the largest schemes to those that develop into cohesive subgroups. Although components of various sizes are present, this does not indicate cohesion. Those involved in the same network component may technically belong to the same group in the whole network, they do not necessarily meet the stricter definition for cohesion established by the Kliquefinder algorithm. Cohesive subgroups are smaller units where individuals are more likely to participate in the same schemes and less likely to participate in schemes outside of their cluster. These subgroups have 109 important characteristics beyond what can be explained by examining network components. The next section details the results of the Kliquefinder analysis. Results of Cohesive Subgroup Analysis Given the separation of the data into distinct components of various sizes, applying Kliquefinder was a unique challenge. The majority of schemes and suspects were removed because they belonged to isolated components not connected to any larger network components. Those with only one connection in the network were removed. That is, individual suspects involved in a single scheme or schemes involving only suspects not linked to other schemes are all removed. The involvement in a single scheme that is unconnected to other parts of the network disqualifies these actors from engaging in cohesive subgroups. Of the remaining nondisconnected actors, a total of 16 suspects and 12 schemes were identified as part of six distinct clusters.14 These clusters are considered cohesive subgroups for the purposes of this study. The results indicated a θ1* value of 4.882 with a p-value of 0.000, indicating that the null hypothesis that θ1 is zero can be rejected. This demonstrates that actors are engaged in schemes together within the identified subgroups in a way that is unlikely to have occurred by chance alone. Monte Carlo simulations of random networks to create a distribution for testing statistical significance were determined to be unnecessary due to the complete separation of components. Simulated distributions would be created using the remaining actors after removing isolated components and those with only one other connection. Visually, the distinctions between these clusters are clearly evident. Therefore, Monte Carlo simulations were not produced. 14 Originally, Kliquefinder clustered together three distinct clusters due to the resistance of the algorithm to creating subgroups of four actors or less. However, further review of the data indicated that these are indeed separate clusters, so this cluster was manually separated into its individual components resulting in the six clusters reported here. 110 Due to the composition and structure of the resulting clusters, MDS was not an ideal technique for visualizing the clusters. As the clusters are entirely separate, MDS places the clusters in the same positions, making distinguishing nodes and edges impossible without manually separating them, thus defeating the purpose of MDS. Therefore, spring-embedded layout was used instead to create the network maps in this chapter. Spring-embedding determines node positions based the geodesic path, where those with the shortest path lengths to each other are placed closer together (Hanneman & Riddle, 2005). Cohesive Subgroup Descriptions This section provides a basic overview of the composition of the clusters. Figures 6.5 and 6.6 illustrate these six subgroups and any additional schemes or suspects they are linked to. The same rules for node/edge color, shape, and size applied above are followed here as well. Each subgroup is circled and identified by number for descriptive purposes (numbers are not assigned in any particular order and do not have substantive meaning). The most common structure (all but one subgroup) is two schemes and two suspects per subgroup. The remaining subgroup (3) contains two schemes and six suspects. The commonalities between these subgroups are that multiple suspects must be involved in numerous scheme activities together, while not being involved in scheme activities with others. This is why many of these subgroups are attached to others in the network not included in the subgroup, where these individuals only participated in a single scheme in the subgroup but not in the other. There are several other similarities between these subgroups. Four are married couples involved in multiple schemes together (subgroups 1, 4, 5, and 6). Two are related to the Institute for Global Prosperity (IGP) organization (subgroups 1 and 3), which is discussed further in Chapter 7. The following is a brief overview of each subgroup. 111 Figure 6.5. Network Sociograms of Cohesive Subgroups Illustrating Ideological Motivation in Far-Right Criminal Anti-Tax Movement: 1990-2010 112 Figure 6.6. Network Sociograms of Cohesive Subgroups Illustrating Greed Motivation in Far-Right Criminal Anti-Tax Movement: 1990-2010 113 Subgroup 4 is a married couple engaged in separate but related tax avoidance schemes. One scheme involved the sale of tax shelter packages and fake financial instruments to satisfy tax debts from 1997 to 2005, where the other was for the couple’s failure to file income tax returns or pay income taxes from 1997 to 1999. The male was 40 and the female 34 at the start of the first scheme. Both are extremist tax protesters with mixed motives and medium strength scores for both ideology and low greed strength scores. The tax avoidance packages they sold were developed by another anti-tax extremist involved in similar financial scheme included in the data, but were not part of the same schemes together. Similar to subgroup 4, subgroup 6 is another married couple (also extremist tax protesters) involved in a tax avoidance and fake check scheme. The first scheme is failure to file or pay income taxes from 1988 to 2006. In the second scheme, the couple relied on the “redemption” theory by attempting to draw on a fake treasury accounts with a “bill of exchange” received from the anti-tax group American Rights Litigators (ARL) in 2005 and 2006. ARL was started by an anti-tax extremist involved in unconnected financial schemes. Both were in their 40s at the start of the first scheme and in their 60s when the schemes ended. Unlike the couple from subgroup 1, these were ideologically motivated individuals and schemes high in ideological strength and low in greed strength. Subgroup 5 is a third married couple that engaged in multiple schemes together, both of which involved other people not included in the cluster. They engaged in both a payroll tax scheme and a warehouse banking scheme. Both schemes were designed to hide income and other assets from the government, either through making untaxed payments to avoid payroll taxes or by maintaining secret accounts in an unregulated banking operation and issuing depositor money from a common pool. Both suspects were in their early 40s at scheme start, tied to a Christian 114 Patriots group, high in ideology strength and low in greed strength, and exhibited many of the same anti-tax philosophies and behaviors as the anti-tax couples from the other subgroups. Subgroup 5 motivations differ from the subgroups 4 and 6. The motives for the schemes were mixed, while their individual motives were ideological. This is in contrast with subgroup 6 where scheme motives were ideological and subgroup 4 where scheme and individual motives were mixed. These differences are largely due to these additional non-extremist collaborators motivated by greed. The remaining three subgroups have vast different features from the others due to their links to both anti-tax organizations and the involvement of a large number of additional schemes or suspects not included in the clusters. Subgroup 2 consisted of two sovereign citizens involved in both tax avoidance and investment schemes. The five-suspect, ideologically driven tax avoidance scheme involved the sale of “pure trusts” and anti-tax literature through a business called the “Freedom Education Center” offering strategies for escaping the U.S. tax system. Each of these individuals also began their own separate tax avoidance schemes by refusing to pay income taxes or file tax returns for several years. The second scheme was a profit motivated investment scheme involving only the two suspects in the subgroup. They sold these fake investments across the U.S. as part of a purported high yield bank debenture program promising 120% profits in 90 days. Both schemes ran from 1997 to 1999 when the suspects were in their early 50s. Each suspect had a strong ideological association centered on sovereign and anti-tax beliefs, but motivations for the two primary collaborative schemes differed drastically. The two remaining subgroups were linked to the same overarching anti-tax multi-level marketing operation known as the Institute for Global Prosperity (IGP). This massive anti-tax operation facilitated numerous schemes and individuals to conduct tax avoidance of various 115 types, sizes, compositions, many of which were unconnected through scheme activities but were still related to IGP in some way. The first involved several IGP leaders who broke off to create their own operation known as Anderson Ark and Associates (AAA). Six of these individuals made up subgroup 3: three anti-tax extremists and three non-extremist collaborators. They promoted a number of anti-tax packages, including one tax avoidance scheme called “Tax Magic” and involved ten additional suspects not included in the subgroup. The second scheme involved money laundering with a program called “Look Forward,” where money from customers was secretly transferred by the perpetrators to Costa Rica and withdraw by the customers in the form of debit cards in order to hide the money from the government. Both schemes operated from 1998 to 2001. Suspects ranged in age from 29 to 57 at the start of the first scheme and consisted of one female and five males. While involvement of the three nonextremists in both schemes was profit motivated, the three extremists were involved for mixed motives although their ideological strength was slightly higher than greed strength. Both schemes had mixed motives. Finally, subgroup 1 was also linked to IGP and involved a married couple selling anti-tax packages known as the “Millennium Program” at IGP events as an independent vendor. Like subgroup 4, this couple not only sold tax shelter packages, but also failed to file their own tax returns or pay income taxes as part of a second scheme. The sale of tax shelters also involved an additional suspect with unknown motivations not included in the cluster. Both schemes and suspects were motivated by a combination of ideology and greed, although the ideological strength of the male perpetrator was stronger than that of the female. Both were in their late 40s and early 50s during the schemes, which lasted from 1999 to 2002. 116 Longitudinal Analysis of Cohesive Subgroups The six subgroups described above shared numerous similarities, but also varied temporally. The forthcoming longitudinal examination focuses more in depth on the evolution of these subgroups between 1990 and 2010. Sociograms with active schemes for each year illustrating both ideology and greed motivations can be found in Appendices F and G. Between 1990 through 1996, only one scheme in subgroup 6 is active, while none of the other subgroups activities have begun. This was a long running ideologically motivated tax avoidance scheme committed by a married couple who later became involved with a larger antitax group to carry out an additional scheme, but whose fraudulent activities began over two decades prior. None of the other subgroups begin until 1997, when two additional subgroups took shape: subgroup 4 and subgroup 2. These subgroups are distinctly different in several ways. Subgroup 4 is a married couple involved in two tax avoidance schemes that both began in 1997. Both of these schemes had mixed motivations. Meanwhile, subgroup 4 involved two sovereign citizens and consisted of tax avoidance and investment schemes that also began the same year. However, their tax avoidance scheme involved several other suspects not included in the subgroup. Each of these additional individuals engaged in their own distinct tax avoidance schemes at the same time by failing to file their tax returns or pay income taxes. At the same time, these two suspects engaged in their own individual tax avoidance schemes not found to be part of the subgroup. These other activities did not involve the commission of multiple schemes by the same individuals, therefore not meeting the criteria for a cohesive subgroup. Perhaps the most active years are from 1998 to 2002. In 1998, subgroup 3 began with both tax avoidance and money laundering operation involving numerous suspects, several of whom were greed motivated non-extremists. Ten additional non-extremist suspects in the tax 117 avoidance scheme were not involved in the money laundering activities, and therefore were not considered part of the cohesive subgroup, which consisted of the six suspects involved in both schemes. This is the largest cohesive subgroup, with the same start and end dates for both sets of scheme activities (1998 to 2001). It appeared to be more highly motivated by greed than the remaining subgroups. In 1999, subgroup 1 was fully formed as a married couple simultaneously began two avoidance schemes, one involving the sale of tax avoidance packages and the other their own personal refusal to pay taxes. As mentioned above, subgroup 1 is tangentially related to subgroup 3 through connections to other individuals not included here, but they were not involved in the same scheme activities together. However at the same time, all the financial scheme activities involved in subgroup 2 ended by 2000 following indictments of all the suspects involved. At the same time, one of the suspects from subgroup 2 began an unrelated investment scheme that would last through 2004. Again though, this was not part of the main subgroup activities. Several other changes occurred in 2000 that lasted through 2001. First, subgroup 5 was fully formed as two separate schemes, bank fraud and tax avoidance. Both began at the same time involving two different sets of suspects. Two of the suspects were a married couple involved in both schemes and made up subgroup 5. Second, the suspects in subgroup 4 ended their individual tax avoidance but continued selling tax avoidance packages. While it should be noted that their personal tax avoidance may very well have continued past 1999, there was not sufficient evidence to support this claim, so it was coded as ending at that time. Third, as described above, subgroup 3, the only subgroup with more than four nodes, was entirely finished by the end of 2001. This was followed the next year by the end of subgroup 1. All of these schemes ended following the indictment and arrest of the suspects involved. 118 The remaining subgroups remained intact from the end of 2002 through 2004. In 2005, the remaining scheme from subgroup 4 ended when the couple was indicted for their sale of antitax “trusts.” The same year, the second scheme from subgroup 6 began when the couple involved in tax avoidance since 1988, following years of exchanges with the IRS, tried to pay their tax debts with a fraudulent “bill of exchange” drawing off a non-existent account at the U.S. Treasury. This continued until 2006 when they were indicted for tax evasion and using fraudulent financial instruments. Subgroup 5 was the final subgroup to carry on until 2009, when both sets schemes were ended at the same time, again through government intervention. Implications There are several implications of the subgroup analysis. As noted above, statistical examination of the data for cohesive subgroups using Kliquefinder is challenged due to the structure and composition of the data. However, this shows a great deal about the nature of offending in the anti-tax movement. There are 152 entirely separable components of various sizes. After removing the 69 components with only two nodes (a single offender involved in a single scheme), 83 components remained ranging from three to twenty-nine nodes. The largest component consisted of a single scheme committed by twenty-eight suspects. This was not found to be a cohesive subgroup because of the large number of people involved in a single scheme. More cohesive subgroups are those where individual suspects are more likely to participate in the same schemes together. This is more likely when the same individuals are involved in multiple schemes together, which is not the case with a large scheme involving multiple individuals. Therefore, Kliquefinder will not identify singular schemes like these as cohesive subgroups. On the one hand, this demonstrates the utility of Kliquefinder to identify those participating in multiple common schemes together. However, if simple coding changes were made to classify a 119 single scheme as multiple or two schemes as one, the resulting cluster assignments would be substantially different. This illustrates an important statistical limitation of the use of two-mode data with Kliquefinder, particularly when the network consists of separable components. Regardless, the algorithm performed as expected, providing further evidence for the disconnected, decentralized nature of criminal offending in the anti-tax movement, consisting mostly of lone perpetrators engaged in their own tax avoidance schemes, but with several individuals engaging in multiple common schemes together to form cohesive subgroups, along with a few more extensive, larger criminal operations that are less cohesive. The longitudinal examination of the cohesive subgroups produced several interesting findings. First, those involved in these cohesive subgroups were likely to begin and end both schemes at the same times. With a few exceptions, there were no major changes found in terms of subgroup activities involving scheme activities starting and ending at different times. In those where this was the case, the activities were still closely related. For subgroup 6, the couple had been involved in long-term tax avoidance by failing to pay taxes until 2005 when they tried to pay their tax debts with a fake “bill of exchange,” a common tactic by those subscribing to the “redemption” theory. As noted above, subgroup 4 showed schemes ending at different times, but this may not have actually been the case due to limited information available on the tax avoidance scheme supposedly ending in first in 1999. Given the role of concentrated activities where those in cohesive subgroups are more likely to participate in criminal activities within the subgroup, this is not entirely surprising. This demonstrates the usefulness of the algorithm by finding scheme and suspects closely tied together and influencing one another in ways that explain the behavior of both units of analysis. The criminal activities in subgroups are highly 120 concentrated both by the individuals involved and the type of schemes they are carrying out. This includes common time periods when these criminal activities primarily occurred. An additional finding was the prevalence of married couples involved in multiple schemes together formulating three of the subgroups and part of a fourth. It is logical that married couples would be likely to act together to carry out financial schemes, particularly tax avoidance, with other illegal activities constituting additional schemes occurring either concurrently or afterward. The marriage constitutes a strong bond and connection beyond cooffending that facilitates cohesion, explaining why they are more likely to be found as cohesive subgroups. However, it was somewhat surprising that the majority of subgroups were married couples. This speaks to the need to examine influences and connections beyond co-offending ties to further understand the development of these unique financial schemes. Finally, this chapter demonstrates the utility of the longitudinal study of networks. This longitudinal examination shows differences as opposed to viewing all the years as a static snapshot. The subgroups and the financial crime scheme activities they represent unfold in different ways over time. While examining the data for all years as one set of data is an important step in understanding the whole picture of what crime in the anti-tax movement looks like, it offers a misleading impression of their scale and scope when the temporal context is ignored. Directions for Additional Research Despite these findings and implications, this analysis also has a notable limitation that provided the direction for the forthcoming qualitative analysis in Chapter 7. Several of these subgroups are connected in ways that are not reflected in the data because individual suspects were not involved in the same scheme activities together, but were limited to the same large 121 overarching group. In this case, subgroups 1 and 3 were linked to IGP, either as part of the main schemes or as a vendor or spinoff group. In order to capture this through the statistical analysis, a third level of analysis capturing involvement in the same overall group or operation would be needed, which is beyond the scope of this study. However, it is notable that even though these schemes and suspects were linked to the same larger group, they were not involved in scheme activities together. Therefore, the finding that these individuals are not involved in the same cohesive subgroups together is valid, as increased cohesion would likely necessitate involvement in the same financial crime activities. This highlights the fact that being linked to the same group or organization does not necessitate that offenders co-offend together or even necessarily interact with one another. This is true for other types of network studies of organizations, where it is not assumed that those linked to the same organization are also linked directly to one another. In order to examine these elements more in depth, a qualitative analysis of IGP and its associated schemes and suspects is undertaken in Chapter 7. Conclusion This chapter examined the relational dynamics of financial criminal activities in the farright anti-tax movement. It was first established that a large number of individuals were involved in financial schemes seemingly unconnected to others, while others were involved in schemes with multiple perpetrators. These larger schemes tended to have greater variation in motives as well as different types of individuals involved in carrying them out. It was then determined that six cohesive subgroups were present in the movement where individuals were more likely to participate in the same financial schemes. These subgroups consisted of multiple individuals involved in more than one financial scheme together. These schemes occurred at different periods of time between 1990 and 2010, although those involved in the same subgroup were 122 likely to carry out their scheme activities at the same times. Finally, it was determined that two of the subgroups were linked together, although not through participating in common schemes but through common association with the same anti-tax organization. These two subgroups and their connections to others in the far-right anti-tax movement are examined further in Chapter 7. 123 CHAPTER 7: QUALITATIVE ANALYSIS The qualitative analysis presented in this chapter unfolds in two main sections. First, the original source materials are examined to learn more about the relational dynamics of those involved in the two subgroups found to have a connection in Chapter 6. Second, the original network from Chapter 6 is enhanced by expanding the two subgroups by incorporating the organizational associations identified in the source materials. A longitudinal examination of the updated network is then conducted to examine how network relationships changes over time. Evaluation of Original Source Materials Prior to discussing the results, additional explication of the specific methodological steps taken to conduct the qualitative analysis is appropriate. As subgroups 1 and 3 had an existing relationship not identified earlier in the subgroup analysis but which became apparent following a preliminary review of the source materials, these subgroups and their additional network connections were selected for more in depth study. The original source materials pertaining to these subgroups were identified and uploaded to MAXQDA to carefully examine the available open source information for commonly occurring individuals and group-based activities, as well as how these subgroups and those they are connected to in the network evolved over time. To accomplish this, individuals involved in the subgroups were first coded. Sections of text pertaining to each individual were examined for any additional information on their introduction into financial scheme activities, as well as their links to other individuals. When a new individual was found to have a relationship to the subgroup in some way, the source materials pertaining to that individual were also uploaded to MAXQDA and included in the analysis. As this process unfolded, it was discovered that those involved in the subgroups were part of a vast interconnected network linked together through their association with four main 124 anti-tax organizations: Institute for Global Prosperity (IGP), Anderson’s Ark and Associates (AAA), Investors International (II), and Pinnacle Quest International (PQI). Each of these organizations was then coded in order to identify additional individuals and schemes with some type of connection. Any additional individuals were again coded. A relational coding matrix was then created to determine which individuals and organizations were discussed near one another in the sources. Higher numbers of overlapping codes is an indication that the two codes were discussed within one paragraph of each other. This effort confirmed those individuals who were linked to each of these organizations as well as those involved in similar scheme activities. The next step was to set reasonable parameters on who would be examined further, as the number of possible individuals included family members, business partners, clients, and numerous others with some relationship to the network. In order to maintain continuity as well as a manageable analysis, only those already included in study sample data were examined as part of the qualitative analysis. Although additional individuals linked to these groups but not meeting the inclusion criteria could be added, this is beyond the scope of the current study and would be a fruitful agenda for future research. Through this effort, it was determined that 23 financial schemes and 63 unique individual suspects were associated with the four organizations identified in the qualitative analysis (overview provided in Tables 7.1 and 7.2). These data were used to create an updated network sociogram to examine the evolution of these groups, schemes, and individuals. 125 Table 7.1. Scheme Characteristics for Updated Network Frequency Percentage Min - Max Mean (SD) Size Lone Wolf Multi-Suspect 22 4 18 100.00 18.18 81.82 - - Type Tax Avoidance Investment Money Laundering 22 18 3 1 100.00 81.86 13.64 4.55 - - Motive Ideological Mixed Profit/Greed 22 4 12 6 100.00 18.18 54.55 27.27 - - Strength of Association Ideology Greed - - 0–4 0–3 1.59 (1.33) 1.64 (1.09) A brief overview of the descriptive statistics reveals some interesting contrasts with study universe. The most substantial difference is that those in this network are far more likely to be non-extremist and have mixed or greed motives than the schemes or suspects as a whole. Whereas the descriptive data of the universe show 61% of schemes were ideologically motivated, the percentage is reduced to 19% in this network, while 23% are greed motivated compared to only 5% of all schemes. Schemes with mixed motives nearly doubled from 34% to 59%. Many of these differences have to do with the numerous individuals involved in most of these schemes as only four were committed by lone offenders while the remaining 18 involved multiple perpetrators. As established in Chapter 5, schemes with larger numbers of individuals are more likely to be greed motivated due to the larger number of individuals involved, making them more likely to involve non-extremist collaborators. 126 Table 7.2. Suspect Characteristics for Updated Network Total (n=62) Frequency Percentage Extremist (n=29) Frequency Percentage Non-Extremist (n=33) Frequency Percentage Sex Male Female 62 41 21 100.00 66.13 33.87 29 19 10 100.00 65.52 34.48 33 22 11 100.00 66.67 33.33 Race White Other 45 41 4 100.00 91.11 8.89 25 22 3 100.00 88.00 12.00 20 19 1 100.00 95.00 5.00 Age (at start) Min - Max Mean (SD) 71 26 – 59 43.31 (8.88) - 40 28 – 59 45.10 (8.70) - 31 26 – 58 41.00 (8.70) - Age (at end) Min - Max Mean (SD) 71 32 – 62 48.24 (8.43) - 40 34 – 62 50.10 (8.15) - 31 32 – 61 45.84 (8.28) - 100.00 66.67 14.81 14.81 3.70 29 13 10 6 0 Region West South Northeast Other 56 31 14 10 1 100.00 55.36 25.00 17.86 1.79 27 18 4 4 1 Affiliation Sovereign Citizen Anti-Tax Non-Extremist 62 19 10 33 100.00 30.65 16.13 53.23 - Organization IGP AAA II PQI Multiple 26 17 6 19 7 41.94 27.42 9.68 30.65 11.29 15 4 4 12 5 51.72 13.79 13.79 41.38 17.24 11 13 2 7 2 33.33 39.39 6.06 21.21 6.06 Motive Ideological Mixed Greed Other 78 7 23 42 6 100.00 8.97 29.49 53.85 1.28 42 7 23 12 0 100.00 16.67 54.76 28.57 0.00 36 30 6 100.00 83.33 16.67 - - 100.00 44.83 34.48 20.69 0.00 - Ideology Strength Min - Max Mean (SD) 78 0–4 0.96 (1.09) - 42 1–4 1.79 (0.84) - - - Greed Strength Min - Max Mean (SD) 78 0–4 2.01 (1.17) - 42 0–4 1.50 (0.83) - 36 0–4 2.61 (1.23) - 127 This is indeed the case here, as more suspects are actually non-extremists than extremist (53% to 47%). This increases the greed motive and strength scores while simultaneously driving down the ideological connections. In this case, mean suspect ideological strength is 1.01, down from 2.37, while mean greed strength is 2.04, up from 1.03. Meanwhile, mean scheme ideological strength decreased from 2.77 to 1.50, while mean scheme greed strength increased from 0.55 to 1.73. There are two potential interpretations of these differences, each of which have merit. One is that the schemes in these networks involving affiliates of these anti-tax organizations are indeed more greed motivated than those involved in financial schemes as a whole. However, an important consideration is that with the increased number of individuals involved in these large-scale financial schemes, less information is available on many of the perpetrators. Therefore, it is more difficult to establish individual ideological motivation while giving the appearance that these suspects are driven by greed. Further evidence for this point is in the similar proportions of extremists driven by greed in this network compared to the data as a whole (30% and 31%, respectively). Additional information may shine more light on their individual views than is apparent through the available source information. Several other similarities and differences are present across the main descriptive variables. Scheme type, sex, race, and age are all similar. Scheme types are largely reflective of the entire study universe: mostly tax avoidance with a few investment schemes along with a money-laundering scheme. Suspects are mostly middle-aged white males. However, there are a slightly greater proportion of females and suspects are several years younger on average than in the study universe. Region differs in that primarily region of residence in this network is concentrated in the West, whereas the study universe was disproportionately found in the South. Finally, if suspects are extremists, they are more likely to be classified as sovereign citizens, 128 compared to anti-tax extremists. However, this difference is not quite as meaningful given the far greater number of non-extremist collaborators. Lastly, connections to the four anti-tax organizations are also displayed in Table 7.2. The percentages and frequencies for this variable do not add up the total number of unique suspects as the suspects were permitted to belong to more than one organization. This was essential for establishing links between organizations in the network analysis presented in the next section. Overall, 42% of the 62 individual suspects were linked with IGP, 31% with PQI, 27% with AAA, and 10% with II, while 11% were linked with multiple organizations. Differences between extremist and non-extremists across organizations were largely consistent with the exception of AAA, which was affiliated with a larger number of non-extremists than the other three organizations. Having provided a basic overview of the schemes and suspects in this subnetwork established through additional review of the source materials, the next section will incorporate these data into an updated network sociogram for further analysis. Expanding Network Beyond Cohesive Subgroups An initial overview of the data indicates that this new network of schemes and individuals related to the four main organizations identified in the source materials are entirely separable into disconnected components. This is similar to the initial findings in Chapter 6 of the sociogram that included all the schemes and suspects in the study universe. In this case, eleven separate components were identified as shown in Figure 7.1 and 7.2. The rules of node/edge color, shape, and size are the same as in Chapter 6, along with the contrast between ideological and greed motivations in the two graphs. Subgroup 4 members are in the lower left corner while subgroup 3 is in the upper-middle section of the graphs. 129 Figure 7.1. Initial Network Sociogram Illustrating Ideological Motivation 130 Figure 7.2. Initial Network Sociogram Illustrating Greed Motivation 131 As is apparent by examining Figure 7.1, the weakness of this initial network is that each of these separable components is connected through their group associations. This is evident from the examination of the source materials described above. This is because they were not involved in the same scheme activities together, a stricter definition outlined by the ECDB inclusion criteria as opposed to group membership. This section expands the network beyond the schemes and suspects to include group membership and reevaluates changes in network connections over time. In order to accomplish this, each of the four organizations was added to the network and connections were made with those individuals directly linked to that organization. This inductive process effectively enhances the initial network by building on the strict binary associations the two-mode network matrix. In order to maintain the two-mode structure, the organizations were treated as their own separate schemes. This meant that schemes would not be connected to the organizations, but individuals were linked to them instead mimicking links to additional schemes. While this approach has limitations, namely giving the appearance of many preexisting schemes being further removed from the center of activity than is actually the case, this approach allowed for similar analytical and visual techniques to be used, thus maintaining consistency throughout the study. This effort allowed for the creation of new sociograms representing the updated network that includes the four anti-tax organizations represented as green triangles. Figures 7.3 and 7.4 illustrate an expanded and more complete network after adding the organizational connections that maintains the strengths of the two-mode analysis. As with the analysis in Chapter 6, it is important to examine this network longitudinally, as these groups emerged over time with subsequent individuals moving from one set of activities to another, spawning additional schemes and groups that are interconnected but still distinguishable. 132 Figure 7.3. Network Sociograms with Anti-Tax Organizations Illustrating Ideological Motivation: 1993 to 2008 133 Figure 7.4. Network Sociograms with Anti-Tax Organizations Illustrating Greed Motivation: 1993 to 2008 134 Evolution of Network As with the evolution of the cohesive subgroups in Chapter 6, this longitudinal analysis focuses in depth on the development of these four main groups identified in the source materials as well as their corresponding schemes and individual suspects. These activities took place between 1993 and 2008. Expanded graphs for Figures 7.3 and 7.4 outlining the scheme (plus groups) and suspect connections active during each year are found in Appendices H and I. The network originated in 1993 with the start of Investors International (II) involving several key individuals that would later shape the other the other three groups. One of these individuals was identified as part of subgroup 3. While two suspects linked to II were extremists, each of these operations appeared to be greed motivated as opposed to ideological. The leader was a forerunner to IGP and AAA and utilized many similar strategies and techniques to develop a sophisticated multi-level marketing scheme, although there was no direct evidence of the use of anti-tax arguments to promote the schemes. Only three suspects were indicted for scheme activities directly related to II, while the others would eventually branch off into their own investment and tax avoidance schemes by formulating the three remaining groups. Meanwhile unrelated to II, several future vendors for the three groups began their own tax avoidance schemes in 1994 and 1995. One was a mixed motive tax avoidance scheme involving several future vendors for PQI (all extremists) that began with the sale of the “pure trust” in 1994. Several of these individuals would later become involved in subsequent tax avoidance schemes related to PQI. Also in 1995, future vendor for IGP began a separate tax avoidance scheme involving nine individual suspects (four extremists and five non-extremists). However, none these individuals had links to any additional schemes involving other groups and only later became involved with IGP to market their own scheme until IGP’s 2002 collapse. 135 1996 was a pivotal year as several individuals involved with II broke away to form IGP after II’s leader came under federal investigation for II’s multi-level marketing scheme. They began their own investment scheme with similar structure and techniques as well as several tax avoidance schemes stemming from their individual refusal to pay income taxes or file tax returns. By 1997, IGP was fully operational and recruiting associates to join, purchase, and sell memberships in their anti-tax programs. Independent (but still affiliated) vendors not involved in leadership or promotion of IGP’s own investment scheme would join IGP at their offshore conferences to promote their own tax avoidance schemes. The scheme promoted by IGP was based not only on anti-tax frivolous arguments about how to “legally” leave the U.S. financial system and move money to overseas accounts but also utilized sovereign citizen philosophy. This was a sharp departure from II, where evidence for the use of sovereign citizen language was lacking in what appeared to be a purely profit driven operation. However, there was a substantial assortment of those who espoused these beliefs and those who did not and all of the IGP linked schemes had mixed motivations of both ideology and greed. By the end of 1997, II would unofficially come to a close, although it’s main leader was not officially sanctioned until 1999 following an investigation by the Securities and Exchange Commission (SEC). Another key year was 1998, when IGP began to come under scrutiny from the federal government for their scheme activities. This began disagreements within IGP over the direction of the organization, resulting in one of the main leaders and founders splitting off from IGP to form AAA. AAA would operate similarly to IGP in structure and technique, although the content and focus of their programs differed. AAA would focus on religious and moral principles (Christian undertones guiding the philosophy of the organization and its activities) while IGP remained more anti-government, promoting common law and sovereign citizens ideology. Like 136 the IGP, both of these AAA schemes were mixed motives. These are the two schemes and their related perpetrated that formed subgroup 3. New members and vendors continued to join both organizations through 2000 to market their own tax avoidance schemes. Among them was the married couple in Subgroup 1 who promoted the “Millennium” program from 1999 until IGP’s end in 2002 in addition to their failure to pay income taxes or file tax returns. While IGP was still in full force come 2000 with vendors continuing to join and start new schemes, several others were beginning to leave either of their own accord or due to law enforcement intervention. Many of these did not appear to have an ideological motivation and were attributed instead to greed. Other schemes had mixed motivations, but none were entirely ideologically motivated. Among these was one of the early members of IGP who was kicked out of the organization after contacting a lawyer to utilize the Sixteenth Amendment defense only to find out it was invalid and could not be argued in court. After telling several other key leaders of this development, all communications were cut off as they continued to promote IGP without him. Other activities involving future members of PQI occurred in 2000 and 2001. One future PQI member began an unrelated tax avoidance scheme. Meanwhile, several others who had begun their own tax avoidance schemes in the mid-1990s ended these tax avoidance schemes. It should be noted that these schemes may have continued, but there was not enough evidence to confirm this. While not directly related, these scheme activities would serve as a precursor to PQI’s anti-tax investment scheme that would begin in the coming years. Another key period was 2001 and 2002, when IGP and AAA both ended and PQI began. AAA ceased operations in 2001 after the leaders and key members were all indicted in federal court. Several IGP vendors and leaders were also indicted in 2001. While AAA was officially 137 over, at least one employee continued a separate tax avoidance scheme (failure to file or pay taxes) until 2004. IGP vendors increasingly dwindled away, particularly after the remaining IGP leadership was indicted in 2002. This same year, two key leaders of IGP were essential to the development of PQI, which stemmed directly from IGP and used many similar materials, techniques, and anti-tax and sovereign citizen ideological arguments. Although neither of these IGP individuals was involved in the same scheme activities as those running PQI, they were directly implicated in the origins of the organization. By 2003, IGP had officially ended, although trial and appeal for both IGP and AAA continued through the mid-2000s. Two vendors would continue their own separate schemes through 2004. PQI was fully operational as the only remaining organization in 2003 continuing the tradition of anti-tax investment schemes originated by II, IGP, and AAA. It was similar to other others in that it involved both extremists and non-extremists and was primarily a mix of both ideological and greed motives. However, two of the extremists linked to the group were involved in their own individual ideologically motivated tax avoidance schemes. PQI would thrive through the 2000s until coming under greater scrutiny in 2005 and 2006. However, a formal injunction would not be granted against PQI and its main leaders and members until 2008, with an indictment being issued later that year. Although the trial and appeal process for PQI continued through 2014, the organization ceased operations by the conclusion of 2008. Implications This analysis shows the network evolution of those linked to larger anti-tax organizations. It differs from the longitudinal evaluation in Chapter 6 because all the actors are connected through these group associations, as opposed to the cohesive subgroup analysis that examined disconnected network components. It demonstrates how a small number of individuals 138 beginning financial crimes in the early 1990s connected to others who subsequently began running their own anti-tax organizations, attracting numerous other individuals who ran similar but independent schemes. These organizations are unique in that they provided a platform for individuals to promote their own schemes in addition to facilitating the operation of the main schemes central to the organization itself. This confirms the potential spread of financial crime activities reaching beyond the immediate confines of a cohesive subgroup or even a few interconnected schemes. The analysis shows how these connections unfold over time, demonstrating the value of the approach. Even after a scheme has ended and a suspect has disappeared from the network, their presence can still be felt through subsequent scheme activities involving those they were previously linked to either through past schemes or group associations. Viewing the entire network with all years in a single sociogram offers a misleading picture of what is occurring and gives the impression of a far more extensive set of ongoing activity than is actually the case. By looking at each year sequentially, the organizations, schemes, and individuals are allowed to emerge and disappear, providing a more accurate description of the composition and behavior of the network. Another important implication is that these schemes were more heavily motivated by greed compared to all schemes in the study universe linked to the anti-tax movement. The reasons for this are largely due to the scale and scope of these four anti-tax organizations. As stated above, larger, more intricate operations require more individuals with diverse skill sets and roles with various motivations for becoming involved in the schemes. Some are interested in promoting their ideology, while others simply see the opportunity for monetary gain. In these cases, the motivations for the schemes are likely to be mixed for these reasons. This is in sharp contrast with the individual tax avoidance schemes, where greater evidence exists for an 139 ideological motive due to the extensive use of anti-tax arguments and the absence of direct evidence of personal greed. Limitations This analysis has several limitations. One centers on the decision to maintain the twomode network structure. Adding the groups associations as a another scheme somewhat negated the impact of numerous schemes and created the impression that the positions of some schemes and suspects are closer together than they actually are. There are numerous other potential approaches, including adding a third mode or focusing on one mode at a time. For example, schemes could be linked to groups, then suspects linked to groups. Again, these alternatives were avoided because the two-mode structure would have to be abandoned. Another limitation was only including those suspects and schemes coded in the ECDB. This analysis had the potential to grow unwieldy, resulting in the need for boundary conditions. Numerous additional individual connections could be established, including vendors, family members, and business associates, to create an even more expansive network. However, little information was known about many of those actually indicted and included in the network and even less is known about unindicted collaborators. Similar to the limitations of open source data collection generally, access to additional information may alter the results in unforeseen ways. Conclusion This chapter expanded the findings from earlier analyses to further examine the relational dynamics of financial criminal activities in the far-right anti-tax movement specifically linked with four anti-tax organizations. These organizations were uncovered due to their connection linking together two separate cohesive subgroups identified in Chapter 6. These organizations were all interconnected through a small number of individuals involved in different but related 140 financial schemes from 1993 to 2008. The longitudinal examination of the network revealed the evolution of these connections and the spread of different criminal activities throughout the network, illustrating characteristics of the network composition not evident with a static approach using a single snapshot. The next chapter will summarize the major findings of the study and outline numerous implications for theory, methods, and policy. 141 CHAPTER 8: DISCUSSION This chapter summarizes the contributions of the current study while also highlighting a number of implications and future research directions. First, an overview of the major findings from the different components of the study is provided. Second, theoretical, methodological, and policy implications are reviewed. Overview of Results The general finding from the descriptive analysis is that the majority are long-term, lonewolf, ideologically motivated tax avoidance schemes using frivolous anti-tax legal arguments committed by white, male, middle-aged individuals from the Southern region of the United States. However, there are many caveats to this broad, somewhat overly simplistic generalization. A large number of schemes involved multiple perpetrators and either mixed motives or those strongly driven by greed. This was particularly true for multi-suspect schemes that were more likely to include a large number of non-extremist collaborators. Lone wolves were responsible for two-thirds of financial scheme activity linked to the anti-tax movement. These schemes were mostly ideologically motivated tax avoidance, but also involved several other scheme types. Although much smaller in number, several lone wolf schemes were motivated by greed instead of ideology or had a mix between the two motives. The large number of lone wolf schemes as opposed to those involving multiple perpetrators tells an interesting story of how anti-tax proponents engage in financial crime schemes. These schemes generally do not require the reliance on others or a great amount of expertise to carry out. Lone wolf schemes involving tax avoidance, false liens, and fake financial instruments do not require any specialized knowledge other than the anti-tax and sovereign citizen arguments that support them, meaning that anyone with access to basic information can file a false return 142 and stop paying taxes, use a fake check, or file a false lien. This knowledge is easily obtained online or through anti-tax seminars commonly held across the U.S. on a regular basis. Multiple perpetrators committed the remaining (one third of the total) financial schemes linked to the anti-tax movement. Tax avoidance schemes involving the sale of various anti-tax products and packages as a business were likely to involve numerous individuals, as well as investment schemes, such as multi-level marketing schemes that were linked to the four larger anti-tax organizations discussed in Chapter 7. However, these larger schemes were more rare, as those involving more than four individuals were only one tenth of all the schemes in the study universe. Multi-suspect schemes were most likely to involve two or three perpetrators. Financial schemes linked to the anti-tax movement were most likely to be motivated by far-right extremist ideology, typically anti-tax and sovereign citizen philosophy. This is true for both lone offender and multi-suspect schemes. Multi-suspect schemes differ though in that they are more likely to include elements of greed as well, particularly due to the inclusion of nonextremist collaborators. Namely, schemes with a larger number of individuals are more likely to have a greater proportion of non-extremist collaborators compared to extremists. This is due in large part to the unique skills needed to pull off more expansive operations but also the fact that the larger operations are inherently less cohesive. The motivations for individual involvement vary to a greater degree than schemes with only two or three people. Even though multi-suspect schemes are more associated with greed than lone wolves, they are not as likely to be purely greed driven, suggesting elements of both ideology and greed as numerous perpetrators with differing goals and purposes for their criminal activities become involved in more intricate scheme operations. 143 Many schemes and individual perpetrators were linked to others also engaging in financial crime schemes. Indeed, over one fifth of those engaged in lone wolf schemes were involved in additional schemes, many of which were linked to other perpetrators. Among those co-offending with others, six cohesive subgroups were identified where multiple individuals participated in the same schemes together. While these schemes included other individuals and the individual involved in subgroups also engaged in additional schemes, these clusters signified highly cohesive interactions occurring concurrently in time. That is, those involved in these cohesive subgroups were likely to begin and end both schemes at the same times with scheme activities that were very closely intertwined. All but two of these subgroups involved married couples acting together, both independently and with collaborators, to carry out multiple schemes, supporting the finding of the large number of couples involved in financial crime schemes. The subgroup analysis provided further evidence for the disconnected, decentralized nature of criminal offending in the anti-tax movement, consisting mostly of lone perpetrators engaged in their own tax avoidance schemes, but with several individuals engaging in multiple common schemes together to form cohesive subgroups, along with a few more extensive, larger criminal operations that are less cohesive. Two of the cohesive subgroups were linked to larger anti-tax organizations. The schemes and suspects associated with these organizations were uniquely different from those in the broader study universe, particularly lone offender schemes. Those involved in criminal activities related to these organizations were more likely to be motivated by greed as well as work with non-extremists to conduct financial schemes on a larger scale. Although larger organizations are rare, they are likely to have a wider impact than those involving disconnected offenders acting on their own. Both the cohesive subgroups and larger organizational activities unfolded in different 144 patterns over time in ways not originally conceived by examining the whole network at a single point in time. A small number of individuals engaged in financial crime schemes whose impact spread and encouraged others (directly or indirectly) to start their own schemes. This demonstrates the usefulness of the longitudinal approach. A final note is essential regarding the general patterns of financial crime offending in terms of scheme counts over time, which coincides with previous findings of the general decline in lone wolf offending over the course of the 2000s after the extraordinarily high levels of violent attacks in the 1990s (see Gruenewald et al., 2013d; Spaaij, 2010). Active schemes rose throughout the late 1990s and peaked around 2002, only to decline steadily throughout the 2000s to levels roughly similar to those from the early 1990s. While this trajectory may indeed be accurate, particularly for the earlier years, active schemes in the later years are likely undercounted. Due to the complex nature of measuring financial schemes, this pattern is not directly comparable to violent incidents. These schemes took place over prolonged periods of time and counting them is a difficult challenge unless an objective, quantifiable measure is used, such as number of indictments in a given year. However, this measure would not reflect the actual timing of the activities and only the end date. When an indictment is issued, it reflects a number of years where schemes are active. Indictments issued after 2010 will reflect financial scheme activities from the late 2000s, and given the nature of lone wolf tax avoidance in particular, these numbers will increase as future financial crimes are uncovered, investigated, and processing through the criminal justice system. Theoretical Implications This study offers several important theoretical implications. First, while various theories added substantively to the understanding of the issues presented in this study and guided the 145 interpretation of the results, theory was not tested. Again as noted in the Chapter 3, the multidimensional and multi-causal nature of anti-tax belief and behavior warranted a selective and purposive examination of theory, which is reviewed here in the context of the findings of current study. Future studies testing various theories of how financial crime schemes develop and operate in the anti-tax movement will be important for furthering the existing knowledge of opportunities and motivations for offending in the context of the network dynamics of the antitax movement. The overwhelming number of ideologically motivated lone offenders involved in financial schemes disconnected from others fits with theories for how individuals come to embrace extremist beliefs and behaviors on an individual basis. Many exhibited both personal and political grievances and externalized the government as the enemy, which is consistent with several of McCauley and Moskalenko’s (2008, 2011) mechanisms of radicalization. Future studies should further examine these mechanisms more in depth. The role of societal forces in shaping financial crime schemes in the anti-tax movement has important theoretical implications. Several other theories could explain radicalization among lone wolf financial offenders, such as mass society theory (consensus) and resource mobilization (conflict) theory, which both focus on societal shifts that result in individuals externalizing the blame for their circumstances toward an outside reference group, which could break down existing social structures and increases the likelihood of adopting anti-tax beliefs. This study focused specifically on the role of network relational connections in shaping behavior, but does not account for societal conditions. These theories are used to shape the understanding of the offending, but not explored in depth. However, the overwhelming number of lone wolf offenders 146 with similar offending patterns despite their disconnection suggests that far-right social movement theories should be further explored to explain the prevalence of lone wolf schemes. As indicated above, the general profile gleamed from this study is of long-term, lone wolf, ideologically motivated tax avoidance schemes using frivolous anti-tax legal arguments committed primarily by white, middle-aged males. While some of these lone offenders undoubtedly knew or were aware of one another, direct relational associations does not account for the vast scope of similar anti-tax scheme activity across time and space. These schemes are highly similar in structure, composition, strategy, and purpose, and offender characteristics. As previously discussed, other forces are at work shaping the tremendous commonalities among schemes. Various economic, political, or other social forces (as outlined in numerous social movement theories) impacted the anti-government and anti-tax sentiments found in the anti-tax movement leading to ideological financial schemes. Among those engaging in financial schemes with co-offenders, a common tie was with family connections. Many of these two person schemes were married couples engaged in joint tax avoidance, such as failure to file tax returns or pay taxes, but others involved couples selling their own anti-tax products, largely mirroring other types of scheme activities but introducing a familial connection driving their interactions as opposed to co-offending ties. Other family members were also featured prominently as co-offenders. This is consistent with prior findings that individuals are often socialized into extremist beliefs and behaviors based on preexisting relational ties (Sageman, 2004), which is consistent with social learning, differential association, and social bond theories. It also supports the important of social cohesion based on homophily, where like individuals choose to engage in behaviors together and influence another based on shared characteristics (explored further below). These findings speak to the need to examine 147 influences and connections beyond co-offending ties, including preexisting family and friendship connections, to further understand the development of these unique financial schemes. Other collaborations were key to many financial crime schemes. The finding of a large number of non-extremist collaborators and various motivations points to a wide variety of reasons why different individuals are involved in different schemes. More generally, criminal networks by definition revolve around illicit activities, although not all persons are involved to the same extent at the same times, with some in the network never actually violating any criminal laws. Some are more involved in criminal activities than others, while some individuals or groups are more central to the network. In order to take advantage of the opportunity structure to carry out criminal operations, different skills, expertise, and status levels in the network are necessary (Levi, 2008; Morselli & Giguere, 2006). However, this study did not focus on different roles filled by those in financial schemes and only identified those holding extremist beliefs and their non-extremist collaborators. It may or may not be the case that extremists hold different roles in extremist financial crimes than non-extremists. Situational theories that focus on opportunities and motivations to engage in criminal behavior will be useful here to examine of different types of individuals become involved in anti-tax behavior, particularly if there are differences based on individuals motivation or extremist association. Network theories will also help determine why some individuals in a more expansive network engage in anti-tax behaviors and others do not, as well as shaping the understanding of behavioral differences based on local interactions (i.e. network positions). Further theorizing regarding different motivations and reasoning behind different types of offenders holding specific network positions will be an important contribution. 148 While this study identified cohesive subgroups and examined their evolution across multiple points in time, it was not designed to examine the intricacies of social influence and social selection in determining how individuals come to participate in financial schemes or how they begin associating with the far-right anti-tax movement generally. While these theoretical ideas are important for understanding network dynamics, this largely descriptive study could not answer these questions. However, this study established the groundwork for more thorough evaluations of influence and selection. Importantly, this study showed that subgroup processes at work, namely that these subgroups adapted and changed over time with actors participating in different and sometimes simultaneous scheme activities, and these schemes often led directly and indirectly to other schemes involving similar techniques and strategies. It is unknown whether individual offenders sought each other out due to similar characteristics or if some offenders playing a primary role in influencing others over time, but both forces are likely at work. Further attention to influence and selection is an essential theoretical area to develop. Theorizing about spread of anti-tax beliefs and behaviors will be important going forward. Examining how these ideas develop and diffuse widely across time and space is important for understanding how the anti-tax ideology functions. Applying diffusion of innovation theory (Rogers, 2003) will provide interesting insights into tracking the adoption of anti-tax behaviors in different contexts. Also important is the role of the Internet in spreading and sustaining these ideas, which is likely substantial. However, the extent of this is unknown. An important issue in terms of the Internet is whether the distribution of anti-tax ideas is sufficient in itself to provoke anti-tax belief or behavior without predisposition, catalyst event, or relational interaction. 149 There is also the chicken and egg problem: do individuals become radicalized through far-right wing propaganda, such as anti-tax literature, they happen to come across online, or do individuals who already hold anti-government views or who wish to avoid paying taxes seek out those promoting anti-tax arguments to support their preexisting beliefs and behavior? The answer depends on individual, situational, and relational considerations. This study dealt with anti-tax beliefs and behaviors but did not speak to the catalyst events or circumstances for the adoption of extremist beliefs. This study identified indicators of extremist ideology and financial crime behaviors occurring concurrently, but not which came first. Life events or circumstances (turning points) outlined in life course theory or the removal of positive stimulus in general strain theory could offer insight into the origin of these anti-tax beliefs and behaviors. While it appears that the adoption of ideology led to financial crime behaviors, the inverse may also be true in many cases, where the internal desire for personal fulfillment, gain, or retribution led individuals to seek out anti-tax extremist belief system to justify those desires. Different theories could potentially explain the causal order here, such as social learning theory focusing on radicalization into anti-tax beliefs and behaviors based on observations of role models and interactions with others who show them how to carry out financial schemes using similar techniques. Others engaged in tax avoidance for personal gain could adopt various techniques of neutralization (Sykes and Matza, 1957) that justify their behavior, which could include anti-tax beliefs. A final theoretical implication is how cohesive subgroups impact survival. Cohesive subgroups influence members above and beyond scheme and organizational connections. It is then logical to consider whether being a member of a subgroup prolongs the ability to engage in financial crime schemes. In other words, does membership in a cohesive subgroup increase the amount of time that individuals are able to participate in financial schemes prior to ending their 150 involvement (voluntarily or involuntarily) compared to those who are not part of cohesive subgroups? Alternatively, does the opposite actually occur, with subgroups leading to decreased prospects for survival? Coupled with this question is whether establishing connections with other subgroups (bridging ties) increases or decreases the odds of survival. Methodological Implications In addition to implications for theory, this study has a number of methodological implications. The first is the tremendous difficulty in studying far-right extremist financial crime. Conceptually, these unique crimes are difficult to situate in the existing literature, presenting challenges for operationalizing key constructs. As established approaches are limited, EFCDB data used in this study is currently the only available database for studying these crimes. Fortunately, this rich database provides the unique opportunity to answer numerous questions and conduct various types of studies that were not previously possible. That said, studies examining the anti-tax movement using non-archival methods, such as interviews, observations, or surveys, will be fruitful endeavors. As this study demonstrates, social network analysis is an ideal way of examining offenders involved in financial crime schemes due to the focus on relational connections in addition to individual and group attributes. The current study and Belli’s (2011) study demonstrate two different network based approaches using ECDB data. While Belli (2011) focused on connections between individual offenders and was able to include family and business ties, the current study incorporated two units of analysis (scheme and suspect) as opposed to the exclusive focus on the individual level. Both designs have their strengths and limitations, but ECDB data support both types of analysis. The benefits of the two-mode network analysis are the preservation of the duality of different levels and the ability to examine the 151 intersection of individual behavior in the context of criminal events. The additional analyses incorporating associations with an anti-tax organization offers one approach to adding a third unit of analysis. While this study took a multi-pronged approach centered on the visualization and description of sequential network snapshots, other methods can be used to examine dynamic social networks, including statistical modeling and tracking network attributes over time. Statistical modeling includes sophisticated approaches for measuring influence and selection. Influence models focus on time and prior behavior, while selection models have the ability to incorporate both individual and event level attributes. Future studies can build on the mixed methods approach taken here and develop statistical models to examine the relative influence of various factors on anti-tax belief and behaviors. Another approach is to focus on changes in network statistics over time, such as various measures of density, centrality, homophily, transitivity, reciprocity, and clustering. While there is clearly a need for further development of approaches to examine relational change, many advances have been made that will prove useful in future research (Moody, McFarland, & Bender-deMoll, 2005; Snijders, 2005, 2009). This study successfully undertook one approach to addressing these analytical limitations, but many other avenues are possible. Another unique aspect of this study was the “open system” approach to data inclusion and network modeling. Instead of expanding out from the individual or event using an egocentric approach, all the schemes and individuals meeting an established set of criteria were incorporated into the network. Naturally, this created some analytical challenges as network members were disconnected and lacked direct associations, as evidenced by the numerous separable components. However, this approach ensured an accurate and complete representation 152 of known criminal activity in the entire anti-tax movement over an extended period of time. The descriptive analysis permitted a full accounting of the various aspects of these schemes and individuals, including characteristics of attributes and social structure. The longitudinal network analyses focused on two different characteristics of the anti-tax movement: clustering in cohesive subgroups and associations with larger anti-tax organizations. Using this approach, individual offending, offending in small cohesive groups, informal network offending, and larger organizational offending were all examined in the same study. Other methodological implications deal with the examination of motivation to participate in financial crime schemes. Various measures were established to examine the contrast between ideological and greed motives for participating in financial schemes. As previously discussed, the ideological association measurement has been carefully established and is useful for distinguishing between those with strong and weak connections to extremist ideology based on identifiable far-right extremist indicators in open source records. Other motivations have not received the same attention. While a greed scale was developed for this study, further refinement on what constitutes greed will be essential going forward. This includes an interesting problem not addressed here related to inherent greed involved in many anti-tax arguments. While tax avoidance schemes in particular present the appearance of ideological motivation without direct evidence of greed, they may in fact be driven by greed. In other words, is the act of tax avoidance in itself an indication of greed due to the resulting personal financial benefit? These situations were not included as evidence of greed for the purposes of this study; otherwise, all the ideologically motivated individuals would have been mixed. However, further study of these complexities is certainly warranted. Finally, other motives for engaging in criminal activity, such as psychological issues and retaliation, were not focused on in this study and deserve increased 153 focus in future studies. In particular, it would be useful to create measures separating the various elements of ideological motives based on these and other contributing factors. Overall, social network analytical techniques like those utilized in this study offer tremendous potential for the development of empirically derived enforcement and prevention policies and practices. Extremists are inherently connected as they operate in different types of networks, and the assumptions of independence of observations are likely invalid, leading to inaccurate conclusions. This study demonstrates the various patterns to how these networks form and uncovering these patterns will advance our knowledge to either confirm or disconfirm currently held assumptions about the structure and characteristics of the anti-tax movement. Further use of these techniques and strategies will continue to yield fruitful results and inform evidence-based policymaking. Policy Implications Finally, this study has clear policy implications. It is first important to reiterate that financial crimes have received minimal attention compared to other types of crime (e.g. violent incidents). This study has provided further evidence that tax evasion and avoidance involving anti-tax extremists is a problem worthy of additional focus. More generally, EFCDB data have identified over 600 financial schemes committed by far-right extremists in the United States since 1990 (Sullivan et al., forthcoming 2015b) with over 350 of these involving identifiable tax protesters (Sullivan et al., forthcoming 2015a). These schemes involved over 1300 individual offenders engaged in the anti-tax movement indicted for financial crimes in the United States since 1990 resulting in over one billion dollars in government losses. This fiscal impact along is a problem deserving the attention of policymakers and practitioners. Evidence-based knowledge of the causes and consequences of extremist financial crime 154 can enhance the decision-making capabilities regarding policy and practice. While financial crime schemes are an important focus for policy on their own, links to violence increase the urgency of the problem. Financial crimes disrupt and defraud both citizens and governments while simultaneously supporting violent crimes by extremist movements, representing an area of growing concern. Identifying links between financial and violent crime activity will be important for policymakers, given the inherent threats to public safety. While not specifically focused on in this study, the ECDB has uncovered tax protesters who escalate from tax avoidance to violence against police officers, other government officials, and the general public. Despite these dangers, fewer resources have been devoted toward far-right domestic terrorism than to international terrorism, and the financial and organizational aspects of extremism in particular have not received adequate focus. Future studies examining connections between financial and violent offenders and contrasting different types of lone wolf offenders will yield important findings. The findings of numerous types of anti-tax offending, ranging from lone offender schemes to multi-suspect schemes as well as the involvement of anti-tax organizations, speaks to the vast nature of the problem and the importance of tailoring responses accordingly. As noted in the review of the white-collar crime literature, those from all areas of society are involved in these financial crimes, ranging from higher (e.g. bank executives, medical professionals) to lower socioeconomic status. There is also the collaboration between so-called legitimate business and illegitimate criminal enterprise, such as when those involved in tax preparation, accounting or legal professional businesses begin to engage in anti-tax behaviors or when they join forces with an anti-tax organization. While those involved with these larger organizations are likely to be most visible and will come under greater scrutiny to law enforcement due to their disproportionate impact, the same strategies may not work as well with lone wolf offenders. 155 However, examining the scope of these larger organizations and other anti-tax promoters seen as leaders in the anti-tax movement could lead directly or indirectly to others who are engaged in their own lone wolf schemes based on the information or advice they received from others. Still others may be receiving their information from the Internet, which requires different strategies from those engaging in anti-tax behaviors based on informal interactions, business relationships, or criminal collaborations. It is important to target these indirect as well as direct networks because they can serve a key role in the spread of anti-tax behaviors. Strategies for policy and practice should be tailored toward the specific type of offending involved and account for potential connections between these offending types. Developing appropriate responses is made more complex by the collaboration of political and religious extremists with non-ideological co-conspirators, as extremists have been shown to engage with purely profit-motivated offenders. For those ideologically motivated by the perceived illegitimacy of the tax system and personal sovereignty from government authority, strategies for appealing to moral consciousness may not be as effective. Associates may be more susceptible to non-traditional interventions including outreach efforts, while tax protesters may be more visible and thus more suitable for traditional criminal justice tactics because they generally do not fear reprisals. Enforcement efforts should focus on everyone involved in criminal networks regardless of motive, expanding beyond those with identifiable extremist links in order to identify those who may be susceptible to adopting extreme anti-tax beliefs or behaviors in the future. It may be fruitful to debrief others who are linked to criminal networks in other ways, including family members, friends, business associates, and community members to minimize potential future offenders by softening hardline views of the government generally and the IRS specifically. 156 As criminal behavior occurs in networks of various type and size, accounting for these network effects is crucial when designing intervention strategies (Andresen & Felson, 2010). Subgroup clusters are formed through different types of ties, with some being more insulated from external intervention than others (Malm et al., 2010). If an individual offender in the network is participating with others and their role is not important to the operation, that person is easily replaceable. Those with a large number of weaker ties to others outside of their subgroups may make ideal influential targets, both to find new information about the patterning of the network and to intervene effectively (Granovetter, 1973). Focusing attention on these bridging ties between different subgroup clusters of the network can be particularly effective at dismantling the entire network (McGloin, 2005a). Understanding network dynamics through a problem-oriented approach is important for intelligence gathering and information sharing across agencies. The network oriented approach taken for this study is consistent with the principles of intelligence-led policing (ILP) (Ratcliffe, 2008). Systematically gathering information on known offenses from multiple sources and using verifiable statistical techniques to analyze patterns in the data has shown tremendous potential for addressing complex crime problems, including criminal activities involving extremists. By focusing efforts on specific crimes or groups through both general patterns and case-specific intelligence, interventions can be specialized to maximize effectiveness (McGarrell, Chermak, & Freilich, 2007), such as with directed police patrols of hot spots (McGarrell, Chermak, Weiss, & Wilson, 2001). Similar approaches can be taken with financial crimes involving tax protesters. Finally, in addition to enforcement and prosecution, interventions not relying on formal criminal justice tactics are vital to crime prevention. The vast array of similar scheme activity across time and space committed by non-associated individuals (including the large number of 157 lone wolf offenders) points to the need for unique policies for increasing tax compliance and addressing anti-tax and anti-government beliefs and attitudes on a wide scale. Soft crime prevention strategies such as those proposed by Belli and Freilich (2009) are an alternative to traditional strategies for tax compliance. These strategies would not replace law enforcement efforts, but are aimed at reducing the overall prominence and acceptance of tax avoidance. Three strategies in particular are important here: (1) neutralization of negative influencers on criminal behavior, (2) introduction of positive role models to counter the lure of anti-tax sentiment, and (3) incentivizing compliance. Altering relational connections by countering the anti-tax message, removing key individuals from the networks, and preventing bridges among subgroups are key components to preventing extremist financial crimes. Conclusion Systematic research on this far-right extremist financial crime is virtually nonexistent, illustrating the contributions of the current study to enhance the existing knowledge of tax protesters (and far-right extremism generally), financial crime schemes, and criminal networks. This study used various methods to examine the network dynamics of financial crime schemes involving individual criminal offenders linked to the far-right extremist anti-tax movement. While not without limitations, this study has made an important contribution to a number of areas, including financial crime, far-right extremism, tax avoidance, motivation and opportunity for crime events, and network dynamics. Future research could move in a number of fruitful directions to expand and improve upon what has been done here, many of which are outlined above. This study has laid the groundwork for an ongoing research agenda using similar methods and theories to enhance the understanding of the various complexities of financial crime behavior involving ideologically and non-ideologically motivated offenders. 158 APPENDICES 159 Appendix A Sources Reviewed by the ECDB to Identify Illegal Violent Incidents and Financial Schemes Official sources Congressional hearing reports Congressional Research Service reports Department of Homeland Security domestic terrorism newsletters and reports Department of Justice agencies press releases about (and sometimes links to) indictments, and convictions about a range of illegal acts FBI's Terrorism in the United States annual report (until 2005) Internal Revenue Service (IRS) website National Counterterrorism Center’s Worldwide Incidents Tracking System (WITS), since 2005 New York State Intelligence Center Reports State and Local Anti-Terrorism Training’s (SLATT) chronology Academic/Scholarly/Think-Tank/Media Databases/Chronologies/Narratives American Terrorism Study (ATS) Avlon’s listing of 45 foiled terror plots since 9/11 (The Daily Best) Bipartisan Policy Center's National Security Preparedness Group (NSPG) reports Brennan Center for Justice at New York University School of Law reports Crenshaw post 9/11 foiled plots chronology Dahl’s listing of unsuccessful plots and attacks against American targets Foundation for Defense of Democracies listings of U.S. terrorist events Global Terrorism Database Heritage Foundation’s listing of 30 foiled terrorist plots post 9/11 160 Hewitt’s Chronology Human Rights First, the NYU Center on Law and Security Institute for Homeland Security Solutions (IHSS): Building on Clues: Examining Successes and Failures in Detecting U.S. Terrorist Plots, 1999- 2009 Kurzman at Triangle Institute for Terrorism and Homeland Security: Muslim-American Terrorism Since 9/11: An Accounting Kushner’s Chronology Mueller’s listing of post 9/11 terrorist plots Monterey Institute’s database on chemical, biological, nuclear cases Mother Jones Profiles in Terror listing alleged domestic terrorists New American Foundation and the Maxwell School, Syracuse University’s listing of extremist suspects that were either indicted or killed between 2001 and 2011 NEFA Foundation, Investigative Project RAND-Memorial Institute for the Prevention of Terrorism’s Terrorism Knowledge Rand Reports by Jenkins & others Shanzer, Kurzman and Moosa NIJ Report: Anti-Terror Lessons of Muslim-Americans Literature review of over 500 articles on far-right, econ/animal rights extremists and Islamic extremists We also conducted systematic media searches of the main LexisNexis interface, the New York Times and other web-sources. Watch-group and movement organizations Animal Liberation Front, Complete U.S. Diary of Actions: The First 30 Years Anti-Defamation League (ADL) chronologies, reports & website (i.e., multiple publications) Arnold Chronology Bryan Denson Incidents Center for Democratic Renewal publications & chronologies 161 Eco Crimes list Fur Commission (news articles summarized) The Foundation for Biomedical Research Illegal Incidents Chronology Jihad Watch website National Animal Interest Alliance website Militia Watchdog Organization Muslim Public Affairs Council: Policy Report Data on Post-9/11 Terrorism in the United States Political Resource Associates publications & chronologies Quatloos, Bombs Taxes & Crayons, The Tax Prophet, and the Tax Protester Dossier for far rightrelated schemes Rick Ross Website Southern Poverty Law Center (SPLC) chronologies, Intelligence Reports & website (i.e., multiple publications) 162 Appendix B Open Source Searching Protocol SEARCHING CASES Each potential identified scheme was treated as a case study with the goal of compiling as much open source information as possible. Each case was systematically searched in existing terrorism databases, official sources, watch-group reports, as well as twenty-six web-engines grouped within a primary and secondary open-source search.15 These searches uncover all published open source materials on each case. Additional criminal cases uncovered during these searches were treated as separate schemes and added to the database. The information uncovered includes media accounts, government documents, court records- indictments, appeals, videos, blogs, books, watch-group reports, movement produced materials, and scholarly accounts. The primary open source search accesses the following seven resources: (1) Lexis-Nexis (2) Proquest (3) Yahoo (4) Google (5) Copernic (6) News Library (7) Westlaw The secondary open source search accesses the following resources: (8) Google Scholar 15 From March 2006 to March 2009, a 27th search engine- Infotrac- was also searched. This engine was then removed from the JJC & MSU online libraries. Infotrac focused on health issues & was used for cases that implicated chemical, biological, nuclear, or radiological weapons. 163 (9) Amazon (10) Google U.S. Government (11) Federation of American Scientists (12) Google Video (13) Center for the Study of Intelligence (14) Surf Wax (15) Dogpile (16) Mamma (17) Librarians’ Internet Index (18) Scirus (19) All the Web (20) Google News (21) Google Blog (22) Homeland Security Digital Library Coders (see below) searched each suspect in four additional search engines to uncover prior and/or subsequent crimes they may have committed: (23) Vinelink, (24) The inmate locator (25) Individual State Department of Corrections (DOCs) (26) Blackbookonline.info 164 Appendix C Ranking of Source Reliability Appellate court proceedings Court proceedings subject to cross examination (e.g., trial transcripts) Court proceedings or documents not subject to cross examination (e.g., indictments) Corroborated information from people with direct access to information provided (e.g., law enforcement and other key informants) Uncorroborated statements from people with that access Media reports Watch-group reports Personal views expressed in blogs, websites, editorials or Op-Ed, etc. 165 Appendix D Strength of Ideological Association Category Criteria Scheme Suspect 4 = Undisputed established ideological motive to further far-right extremist goals 4 = Undisputed established present or past adherence to far-right extremist ideology 1) Multiple (2 or more) far-right extremist indicators found, and 2) No evidence found contrary to far-right extremist association 3 = Clear established ideological motive to further far‐right extremist goals 3 = Clear established present or past adherence to far‐right extremist ideology 1) Only single far‐right extremist indicator found, and 2) No evidence found contrary to far-right extremist association 2 = Disputed established ideological motive to further far‐right extremist goals 2 = Disputed present or past adherence to far‐right extremist ideology 1) Multiple (2 or more) far-right extremist indicators found, and 2) Evidence found contrary to far‐right extremist association 1 = Disputed established ideological motive to further far‐right extremist goals 1 = Disputed established present or past adherence to far‐right extremist ideology 1) Only single far‐right extremist indicator found, and 2) Evidence found contrary to far-right extremist association 0 = No apparent ideological motive 0 = No established present or past adherence to farright extremist ideology 1) No far-right extremist indicator found, and 2) Evidence found contrary to far-right extremist association 166 Appendix E Strength of Greed Motive Category Scheme Criteria Suspect 4 = Undisputed established greed motive; no apparent ideological motive 4 = Undisputed established greed motive; no apparent connection to extremist ideology 1) Multiple (2 or more) greed indicators found, and 2) No evidence found contrary to establish another motive (ideology, etc.) 3 = Clear established greed motive 3 = Clear established greed motive 1) Only single greed indicator found, and 2) No evidence found contrary to establish another motive (ideology, etc.) 2 = Disputed established greed motive; evidence of other potential motives (e.g. ideology, etc.) 2 = Disputed greed motive; evidence of other potential motives (e.g. ideology, mental illness etc.) 1) Multiple (2 or more) greed indicators found, and 2) Evidence found contrary to greed motive (indicators of extremist ideology found) 1 = Disputed greed motive; established evidence of other motives (e.g. ideology, etc.) 1 = Disputed greed motive; established evidence of other motives (e.g. ideology, mental illness etc.) 1) Only single greed motive indicator found, and 2) Evidence found contrary to greed motive (indicators of extremist ideology found) 0 = No established greed motive 0 = No established greed motive 1) No greed indicator found, and 2) Evidence found contrary to greed motive (evidence of extremist ideology, etc.) 167 Appendix F Figure A.1. Network Evolution Sociograms of Cohesive Subgroups Illustrating Ideological Motivation in Far-Right Criminal Anti-Tax Movement These graphs represent the longitudinal analysis of the cohesive subgroups in the far-right antitax movement as determined by Kliquefinder. These sociograms represent ideological motivations. Larger scheme nodes indicate stronger scheme ideological strength and smaller nodes indicate low scheme ideological strength (scale “0” to “4”). Similarly, thicker edges represent stronger suspect ideological strength, and thinner edges indicate low suspect ideological strength (scale “0” to “4”). Snapshots of active scheme activities for each year are represented through separate sociograms. Years with no changes are combined in the same graph. Node positions are held constant while and edges are activated and deactivated depending on whether or not the scheme was active in that particular year. The results illustrate the dynamic network change from year to year to obtain a temporal picture of how the subgroups unfold over time. 168 Figure A.1. Network Evolution Sociograms of Cohesive Subgroups Illustrating Ideological Motivation in Far-Right Criminal Anti-Tax Movement Years: 1990-1996 169 Figure A.1 (cont’d) Year: 1997 170 Figure A.1 (cont’d) Year: 1998 171 Figure A.1 (cont’d) Year: 1999 172 Figure A.1 (cont’d) Years: 2000-2001 173 Figure A.1 (cont’d) Year: 2002 174 Figure A.1 (cont’d) Years: 2003-2004 175 Figure A.1 (cont’d) Year: 2005 176 Figure A.1 (cont’d) Year: 2006 177 Figure A.1 (cont’d) Year: 2007-2009 178 Appendix G Figure A.2. Network Evolution Sociograms of Cohesive Subgroups Illustrating Greed Motivation in Far-Right Criminal Anti-Tax Movement These graphs represent the longitudinal analysis of the cohesive subgroups in the far-right antitax movement as determined by Kliquefinder. These sociograms represent greed motivations. Larger scheme nodes indicate stronger scheme greed strength and smaller nodes indicate low scheme greed strength (scale “0” to “4”). Similarly, thicker edges represent stronger suspect ideological strength, and thinner edges indicate low suspect ideological strength (scale “0” to “4”). Snapshots of active scheme activities for each year are represented through separate sociograms. Years with no changes are combined in the same graph. Node positions are held constant while and edges are activated and deactivated depending on whether or not the scheme was active in that particular year. The results illustrate the dynamic network change from year to year to obtain a temporal picture of how the subgroups unfold over time. 179 Figure A.2. Network Evolution Sociograms of Cohesive Subgroups Illustrating Greed Motivation in Far-Right Criminal AntiTax Movement Years: 1990-1996 180 Figure A.2 (cont’d) Year: 1997 181 Figure A.2 (cont’d) Year: 1998 182 Figure A.2 (cont’d) Year: 1999 183 Figure A.2 (cont’d) Years: 2000-2001 184 Figure A.2 (cont’d) Year: 2002 185 Figure A.2 (cont’d) Years: 2003-2004 186 Figure A.2 (cont’d) Year: 2005 187 Figure A.2 (cont’d) Year: 2006 188 Figure A.2 (cont’d) Years: 2007-2009 189 Appendix H Figure A.3. Network Evolution Sociograms with Anti-Tax Organizations Illustrating Ideological Motivation These graphs represent the longitudinal analysis of the updated network with the four anti-tax organizations in a way that mimics the addition of four schemes. The four organizations are labeled and presented as green triangles. These sociograms represent ideological motivations. Larger scheme nodes indicate stronger scheme ideological strength and smaller nodes indicate low scheme ideological strength (scale “0” to “4”). Similarly, thicker edges represent stronger suspect ideological strength, and thinner edges indicate low suspect ideological strength (scale “0” to “4”). Snapshots of active scheme activities for each year are represented through separate sociograms. Years with no changes are combined in the same graph. Node positions are held constant while and edges are activated and deactivated depending on whether or not the scheme or organization was active in that particular year. The results illustrate the dynamic network change from year to year to obtain a temporal picture of how the network unfolds over time. 190 Figure A.3. Network Evolution Sociograms with Anti-Tax Organizations Illustrating Ideological Motivation Year: 1993 191 Figure A.3 (cont’d) Year: 1994 192 Figure A.3 (cont’d) Year: 1995 193 Figure A.3 (cont’d) Year: 1996 194 Figure A.3 (cont’d) Year: 1997 195 Figure A.3 (cont’d) Year: 1998 196 Figure A.3 (cont’d) Year: 1999 197 Figure A.3 (cont’d) Year: 2000 198 Figure A.3 (cont’d) Year: 2001 199 Figure A.3 (cont’d) Year: 2002 200 Figure A.3 (cont’d) Year: 2003 201 Figure A.3 (cont’d) Year: 2004 202 Figure A.3 (cont’d) Year: 2005 203 Figure A.3 (cont’d) Year: 2006 204 Figure A.3 (cont’d) Years: 2007-2008 205 Appendix I Figure A.4. Network Evolution Sociograms with Anti-Tax Organizations Illustrating Greed Motivation These graphs represent the longitudinal analysis of the updated network with the four anti-tax organizations added as schemes. These sociograms represent greed motivations. Larger scheme nodes indicate stronger scheme greed strength and smaller nodes indicate low scheme greed strength (scale “0” to “4”). Similarly, thicker edges represent stronger suspect greed strength, and thinner edges indicate low suspect greed strength (scale “0” to “4”). Snapshots of active scheme activities for each year are represented through separate sociograms. Years with no changes are combined in the same graph. Node positions are held constant while and edges are activated and deactivated depending on whether or not the scheme or organization was active in that particular year. The results illustrate the dynamic network change from year to year to obtain a temporal picture of how the network unfolds over time. 206 Figure A.4. Network Evolution Sociograms with Anti-Tax Organizations Illustrating Greed Motivation Year: 1993 207 Figure A.4 (cont’d) Year: 1994 208 Figure A.4 (cont’d) Year: 1995 209 Figure A.4 (cont’d) Year: 1996 210 Figure A.4 (cont’d) Year: 1997 211 Figure A.4 (cont’d) Year: 1998 212 Figure A.4 (cont’d) Year: 1999 213 Figure A.4 (cont’d) Year: 2000 214 Figure A.4 (cont’d) Year: 2001 215 Figure A.4 (cont’d) Year: 2002 216 Figure A.4 (cont’d) Year: 2003 217 Figure A.4 (cont’d) Year: 2004 218 Figure A.4 (cont’d) Year: 2005 219 Figure A.4 (cont’d) Year: 2006 220 Figure A.4 (cont’d) Year: 2007-2008 221 REFERENCES 222 REFERENCES ADL. (2005). Extremism in America: Tax protest movement. Washington DC: Anti-Defamation League. Retrieved from http://www.adl.org/learn/ext_us/TPM.asp?xpicked=4&item=21 ADL. (2012). The lawless ones: The resurgence of the sovereign citizen movement (2nd ed.). Washington DC: Anti-Defamation League. Retrieved from http://www.adl.org/assets/pdf/combating-hate/Lawless-Ones-2012-Edition-WEBfinal.pdf Agnew, R. (1992). Foundation for a general strain theory of crime and delinquency. Criminology, 30(1), 47-87. Agnew, R. (2010). A general strain theory of terrorism. Theoretical Criminology, 14(2), 131153. Aho, J. A. (1990). The politics of righteousness: Idaho Christian patriotism. Seattle: University of Washington Press. Akers, R. L. (1973). Deviant behavior: A social learning approach. Belmont, CA: Wadsworth. Akers, R. L. (2011). Social learning and social structure: A general theory of crime and deviance. Piscataway, NJ: Transaction Publishers. Albanese, J. S. (2000). The causes of organized crime: Do criminal organize around opportunities for crime or do criminal opportunities create new offenders? Journal of Contemporary Criminal Justice, 16(4), 409-423. Albanese, J. S. (2011). Transnational crime and the 21st century. New York: Oxford University Press. Alexander, M. C., & Danowski, J. A. (1990). Analysis of an ancient network: Personal communication and the study of social structure in a past society. Social Networks, 12(4), 313-335. Andrei, A. L., Comer, K., & Koehler, M. (2014). An agent-based model of network effects on tax compliance and evasion. Journal of Economic Psychology: Special Issue on Behavioral Dynamics of Tax Evasion, 40, 119–133. Andresen, M. A., & Felson, M. (2010). The impact of co-offending. British Journal of Criminology, 50(1), 66-81. Asal, V., & Rethemeyer, R. K. (2006). Researching terrorist networks. Journal of Security Education, 1(4), 65-74. Baker, W. E., & Faulkner, R. R. (1993). The social organization of conspiracy: Illegal networks in the Heavy Electrical Equipment Industry. American Sociological Review, 58(6), 837860. 223 Baker, W. E., & Faulkner, R. R. (2003). Diffusion of fraud: Intermediate economic crime and investor dynamics. Criminology, 41(4), 1173-1206. Bakker, E. (2006). Jihadi terrorists in Europe, their characteristics and the circumstances in which they joined the jihad: An exploratory study. The Hague: Clingendael Institute. Barkun, M. (1997). Religion and the racist right: The origins of the Christian Identity movement. Chapel Hill, NC: University of North Carolina Press. Beare, M. (Ed.). (2003). Critical reflections on transnational organized crime, money laundering and corruption. Toronto: Toronto University Press. Belli, R. (2011). Where political extremist and greedy criminal meet: A comparative study of financial crimes and criminal networks in the United States. (PhD dissertation), John Jay College, The Graduate Center, City University of New York, New York. Retrieved from http://www.ncjrs.gov/App/Publications/abstract.aspx?ID=256482 Belli, R., & Freilich, J. D. (2009). Situational crime prevention and non-violent terrorism: A “soft” approach against ideologically motivated tax-refusal. Crime Prevention Studies, 25, 173-206. Benson, M. L., & Simpson, S. S. (2009). White-collar crime: An opportunity perspective. New York: Routledge. Blee, K. M. (2003). Inside organized racism: Women and men in the hate movement. Berkeley: University of California Press. Block, A. A. (Ed.). (1991). The business of crime: A documentary study of organized crime in the American economy. Oxford: Westview Press. Borgatti, S. P., & Everett, M. G. (1997). Network analysis of 2-mode data. Social Networks, 19(3), 243-269. Borgatti, S. P., Everett, M. G., & Freeman, L. (2002). UCINET for Windows: Software for social network analysis. Harvard, MA: Analytic Technologies. Retrieved from http://www.analytictech.com/ucinet/ Braithwaite, J. (2001). Conceptualizing organizational crime in a world of plural cultures. In H. N. Pontell & D. Shichor (Eds.), Contemporary issues in crime and criminal justice: Essays in honor of Gilbert Geis (pp. 17-32). Upper Saddle River, NJ: Prentice Hall. Braithwaite, V. (2010). Criminal prosecution within responsive regulatory practice. Criminology & Public Policy, 9(3), 515-523. Breiger, R. L. (1974). The duality of persons and groups. Social Forces, 53(2), 181-190. Burt, R. S. (1992). Structural holes: The social structure of competition. Cambridge, MA: Harvard University Press. 224 Burt, R. S. (2005). Brokerage and social closure: An introduction to social capital. New York: Oxford University Press. Byrne, D. (1971). The attraction paradigm (Vol. 11). New York: Academic Press. Byrne, D. (1997). An overview (and underview) of research and theory within the attraction paradigm. Journal of Social and Personal Relationships, 14(3), 417-431. Carrington, P. J. (2011). Crime and social network analysis. In J. Scott & P. J. Carrington (Eds.), The SAGE handbook of social network analysis (pp. 236-270). Thousand Oaks, CA: Sage. Carter, D. L., Chermak, S. M., Carter, J. G., & Drew, J. (2014). Understanding law enforcement intelligence processes. Report to the Resilient Systems Division, Science and Technology Directorate, U.S. Department of Homeland Security. College Park, MD: National Consortium for the Study of Terrorism and Responses to Terrorism (START).. Retrieved from http://www.start.umd.edu/pubs/START_UnderstandingLawEnforcementIntelligenceProc esses_July2014.pdf Caspi, D., Freilich, J. D., & Chermak, S. M. (2012). Worst of the bad: Violent white supremacist groups and lethality. Dynamics of Asymmetric Conflict: Pathways Toward Terrorism and Genocide, 5(1), 1-17. Charmaz, K., & Bryant, A. (2011). Grounded theory and credibility. In D. Silverman (Ed.), Qualitative Research (3rd ed., pp. 291-309). Thousand Oaks, CA: Sage. Chermak, S. M. (2002). Searching for a demon: Media construction of the militia movement. Boston: Northeastern University Press. Chermak, S. M., Freilich, J. D., Parkin, W. S., & Lynch, J. P. (2012). American terrorism and extremist crime data sources and selectivity bias: An investigation focusing on homicide events committed by far-right extremists. Journal of Quantitative Criminology, 28(1), 191-218. Chermak, S. M., Freilich, J. D., & Shemtob, Z. (2009). Law enforcement training and the domestic far right. Criminal Justice and Behavior, 36(12), 1305-1322. Chermak, S. M., Freilich, J. D., & Simone, J. (2010). Surveying American state police agencies about lone wolves, far-right criminality, and far-right and Islamic jihadist criminal collaboration. Studies in Conflict & Terrorism, 33(11), 1019-1041. Chermak, S. M., Freilich, J. D., & Suttmoeller, M. (2013). The organizational dynamics of farright hate groups in the United States: Comparing violent to nonviolent organizations. Studies in Conflict & Terrorism, 36(3), 193-218. Chermak, S. M., & Gruenewald, J. A. (2006). The media's coverage of domestic terrorism. Justice Quarterly, 23(4), 428-461. 225 Coles, N. (2001). It's not what you know - It's who you know that counts: Analysing serious crime groups as social networks. British Journal of Criminology, 41(4), 580-594. Corcoran, J. (1990). Bitter harvest, Gordon Kahl, and the Posse Comitatus: Murder in the heartland. New York: Penguin. Dahl, E. J. (2011). The plots that failed: Intelligence lessons learned from unsuccessful terrorist attacks against the United States. Studies in Conflict & Terrorism, 34(8), 621-648. Damphousse, K. R. (2010). Recommendations about recommendations: Regarding the need for sufficient funding, sophisticated data analysis, and discipline maturity. In N. A. Frost, J. D. Freilich, & T. R. Clear (Eds.), Contemporary issues in criminal justice policy: Policy proposals from the American Society of Criminology Conference (pp. 151-157). Belmont, CA: Wadsworth, Cengage Learning. Diani, M. (1992). The concept of a social movement. Sociological Review, 40(1), 1-25. Dyer, J. (1997). Harvest of rage: Why Oklahoma City is only the beginning. Boulder, CO: Westview Press. Ekblom, P., & Tilley, N. (2000). Going equipped: Criminology, situational crime prevention and the resourceful offender. British Journal of Criminology, 40(3), 376-398. Emirbayer, M. (1997). Manifesto for a relational sociology. American Journal of Sociology, 103(2), 281-317. Erickson, B. H. (1988). The relational basis of attitudes. In B. Wellman & S. D. Berkowitz (Eds.), Social structures: A network approach (pp. 99-121). Cambridge, UK: Cambridge University Press. Ezekiel, R. S. (1995). The racist mind: Portraits of neo-Nazis and Klansmen. New York: Viking/Penguin. Ezekiel, R. S. (2002). An ethnographer looks at Neo-Nazi and Klan groups: The Racist Mind revisted. American Behavioral Scientist, 46(1), 51-71. Faulkner, R. R., Cheney, E. R., Fisher, G. A., & Baker, W. E. (2003). Crime by committee: Conspirators and company men in the illegal electrical industry cartel, 1954-1959. Criminology, 41(2), 511-554. FBI. (2007). Strategic plan: 2004-2009. Washington, DC: United States Printing Office. Field, S., Frank, K. A., Schiller, K., Riegle-Crumb, C., & Muller, C. (2006). Identifying positions from affiliation networks: Preserving the duality of people and events. Social Networks, 28(2), 97–123. Flynn, K., & Gerhardt, G. (1990). The silent brotherhood: The chilling inside story of America’s violent anti-government militia movement. New York: Signet/Penguin. 226 Frank, K. A. (1995). Identifying cohesive subgroups. Social Networks, 17(1), 27-56. Frank, K. A. (1996). Mapping interactions within and between cohesive subgroups. Social Networks, 18(2), 93-119. Frank, K. A., & Yasumoto, J. Y. (1996). Embedding subgroups in the sociogram: Linking theory and image. Connections, 19(1), 43-57. Frank, K. A., & Yasumoto, J. Y. (1998). Linking action to social structure within a system: Social capital within and between subgroups. American Journal of Sociology, 104(3), 642-686. Frank, K. A., & Zhao, Y. (2005). Subgroups as a meso-level entity in the social organization of schools. In L. Hedges & B. Schneider (Eds.), Social organization of schools: Book honoring Charles Bidwell’s retirement (pp. 279–318). New York: Sage. Freilich, J. D. (2003). American militias: State-level variations in militia activities. New York: LFB Scholarly Publishing. Freilich, J. D., & Chermak, S. M. (2009). Preventing deadly encounters between law enforcement and American far-rightists. Crime Prevention Studies, 25, 141-172. Freilich, J. D., Chermak, S. M., Belli, R., Gruenewald, J., & Parkin, W. S. (2014). Introducing the United States Extremist Crime Database (ECDB). Terrorism & Political Violence, 26(2), 372-384. Freilich, J. D., Chermak, S. M., & Caspi, D. (2009). Critical events in the life trajectories of domestic extremist groups: A case study analysis of four violent white supremacist groups. Criminology & Public Policy, 8(3), 497-530. Freilich, J. D., Chermak, S. M., Gruenewald, J. A., & Parkin, W. S. (2012). Far-right violence in the United States: 1990-2010. College Park, MD: National Consortium for the Study of Terrorism and Responses to Terrorism (START). Retrieved from http://www.start.umd.edu/sites/default/files/files/publications/br/ECDB_FarRight_FactSh eet.pdf Freilich, J. D., Chermak, S. M., & Simone, J. (2009). Surveying American state police agencies about terrorism threats, terrorism sources, and terrorism definitions. Terrorism and Political Violence, 21(3), 450-475. Freilich, J. D., Pichardo Almanzar, N. A., & Rivera, C. J. (1999). How social movement organizations explicitly and implicitly promote deviant behavior: The case of the militia movement. Justice Quarterly, 16(3), 655-683. Freilich, J. D., & Pridemore, W. A. (2005). A reassessment of state-level covariates of militia groups. Behavioral Sciences and the Law, 23(4), 527-546. 227 Friedkin, N. E. (2006). A structural theory of social influence (Vol. 13). Cambridge: Cambridge University Press. Friedrichs, D. O. (2004). Trusted criminals: White-collar crime in contemporary society (2nd ed.). Belmont, CA: Thomson Wadsworth. Geis, G. (2007). White-collar and corporate crime. Upper Saddle River, NJ: Pearson Prentice Hall. Gibson, J. W. (1994). Warrior dreams: Paramilitary culture in the post Vietnam America. New York: Hill and Wang. Gill, P., Lee, J., Rethemeyer, R. K., Horgan, J., & Asal, V. (2014). Lethal connections: The determinants of network connections in the Provisional Irish Republican Army 19701998. International Interactions: Empirical and Theoretical Research in International Relations, 40(1), 52-78. Granovetter, M. S. (1973). The strength of weak ties. American Journal of Sociology, 78(6), 1360-1380. Griffin, S. P. (2002). Actors or activities? On the social construction of "white-collar crime" in the United States. Crime, Law & Social Change, 37(3), 245-276. Gross, E. (1980). Organizational structure and organizational crime. In G. Geis & E. Stotland (Eds.), White-collar crime: Theory and research (pp. 52-75). Beverly Hill, CA: Sage. Gruenewald, J. A., Chermak, S. M., & Freilich, J. D. (2013a). Distinguishing “loner” attacks from other domestic extremist violence: A comparison of far-right homicide incident and offender characteristics. Criminology & Public Policy, 12(1), 63-91. Gruenewald, J. A., Chermak, S. M., & Freilich, J. D. (2013d). Far-right lone wolf homicides in the United States. Studies in Conflict and Terrorism, 36(12), 1005-1024. Gruenewald, J. A., Freilich, J. D., & Chermak, S. M. (2009). An overview of the domestic farright and its criminal activities. In B. Perry & R. Blazak (Eds.), Hate crimes: Hate crime offenders (pp. 1-21). Westport: Praeger. Halebsky, S. (1976). Mass society and political conflict: Towards a reconstruction of theory. New York: Cambridge University Press. Hamm, M. S. (1994). American skinheads: The criminology and control of hate crime. Westport, CT: Praeger. Hamm, M. S. (2005). Crimes committed by terrorist groups: Theory, research, and prevention. Washington, DC: Final report to the National Institute of Justice, U.S. Department of Justice. Retrieved from https://www.ncjrs.gov/pdffiles1/nij/grants/211203.pdf 228 Hamm, M. S., & van de Voorde, C. (2005). Crimes committed by terrorist groups: Theory, research, and prevention. Trends in Organized Crime, 9(2), 18-50. Hanneman, R. A., & Riddle, M. (2005). Introduction to social network methods. Riverside, CA: University of California, Riverside. Retrieved from http://faculty.ucr.edu/~hanneman/nettext/ Hirschi, T. (1969). Causes of delinquency. Berkeley: University of California Press. Hofstadter, R. (1964). The paranoid style in American politics. New York: Vintage. Hollstein, B. (2011). Qualitative approaches. In J. Scott & P. J. Carrington (Eds.), The SAGE handbook of social network analysis (pp. 404-416). Thousand Oaks, CA: Sage. IRS. (2013). IRS releases the Dirty Dozen Tax Scams for 2013. Washington DC: Internal Revenue Service. Retrieved from http://www.irs.gov/uac/Newsroom/IRS-Releases-theDirty-Dozen-Tax-Scams-for-2013 IRS. (2014). The truth about frivolous tax arguments. Washington, DC: Internal Revenue Service. Retrieved from http://www.irs.gov/pub/irs-utl/friv_tax.pdf Iwanski, N., & Frank, R. (2014). The evolution of a drug co-arrest network. In C. Morselli (Ed.), Crime and networks (pp. 52-80). New York: Routledge. Kandel, D. B. (1978). Homophily, selection, and socialization in adolescent friendships. American Journal of Sociology, 84(2), 427-436. Kaplan, J. (1997). Leaderless resistance. Terrorism & Political Violence, 9(3), 80-95. Katz, J. (1983). A theory of qualitative methodology: The social system of analytic fieldwork. In R. M. Emerson (Ed.), Contemporary field research: A collection of readings (pp. 127148). Boston: Little, Brown & Co. Kennedy, D. M. (2009). Deterrence and crime prevention: Reconsidering the prospect of sanction (Vol. 2). London: Routledge. Kirchler, E. (2007). The economic psychology of tax behavior. Cambridge: Cambridge University Press. Klandermans, B., Staggenborg, S., & Tarrow, S. (2002). Blending methods and building theories in social movement research. In B. Klandermans & S. Staggenborg (Eds.), Methods of social movement research (pp. 314-349). Minneapolis: University of Minnesota Press. Klerks, P. (2002). The network paradigm applied to criminal organisations: Theoretical nitpicking or a relevant doctrine for investigators? Recent developments in the Netherlands. Connections, 24(3), 53-65. 229 Koschade, S. (2006). A social network analysis of Jemaah Islamiyah: The applications to counterterrorism and intelligence. Studies in Conflict & Terrorism, 29(6), 559-575. Krebs, V. E. (2002). Mapping networks of terrorist cells. Connections, 24(3), 43-52. LaFree, G., Dugan, L., Fogg, H., & Scott, J. (2006). Building a global terrorism database: Final report to the National Institute of Justice. Retrieved from https://www.ncjrs.gov/pdffiles1/nij/grants/214260.pdf Leenders, R. T. A. J. (1997). Longitudinal behavior of network structure and actor attributes: Modeling interdependence of contagion and selection. In P. Doreian & F. N. Stokman (Eds.), Evolution of social networks (pp. 165-184). Amsterdam: Gordon & Breach. Leighton, P. (2010). Fairness matters-more than deterrence: Class bias and the limits of deterrence. Criminology & Public Policy, 9(3), 525-533. Leuprecht, C., & Hall, K. (2013). Networks as strategic repertoires: Functional differentiation among Al-Shabaab terror cells. Global Crime, 14(2-3), 287-310. Levi, M. (2003). Organising and controlling payment card fraud: fraudsters and their operational environment. Security Journal, 16(2), 21-30. Levi, M. (2008). Organised fraud: Unpacking research on networks and organisation. Criminology & Criminal Justice, 8(4), 389-420. Levi, M. (2010). Serious tax fraud and noncompliance: A review of evidence on the differential impact of criminal and noncriminal proceedings. Criminology & Public Policy, 9(3), 493513. Levitas, D. (2001). Tracing the opposition to taxes in America. Montgomery, AL. Southern Poverty Law Center. Retrieved from http://www.splcenter.org/get-informed/intelligencereport/browse-all-issues/2001/winter/untaxing-america Lewins, A., & Silver, C. (2007). Using software in qualitiative research: A step-by-step guide. Thousand Oaks, CA: Sage. Lipset, S. M., & Raab, E. (1970). The politics of unreason: Right-wing extremism in America, 1790-1970. New York: Harper Torchbook. Malm, A. E., & Bichler, G. (2011). Networks of collaborating criminals: Assessing the structural vulnerability of drug markets. Journal of Research in Crime and Delinquency, 48(2), 271-297. Malm, A. E., Bichler, G., & Van De Walle, S. (2010). Comparing the ties that bind criminal networks: Is blood thicker than water? Security Journal, 23(1), 52-74. Malm, A. E., Kinney, J. B., & Pollard, N. R. (2008). Social network and distance correlates of criminal associates involved in illicit drug production. Security Journal, 21(1), 77-94. 230 Mason, C. (2002). Killing for life: The apocalyptic narrative of pro-life politics. Ithaca, NY: Cornell University Press. Matthew, R., & Shambaugh, G. (2005). The limits of terrorism: A network perspective. International Studies Review, 7(4), 617-627. McAllister, B. (2004). Al Qaeda and the innovative firm: Demythologizing the network. Studies in Conflict & Terrorism, 27(4), 297-319. McCauley, C., & Moskalenko, S. (2008). Mechanisms of political radicalization: Pathways toward terrorism. Terrorism & Political Violence, 20(3), 415-433. McCauley, C., & Moskalenko, S. (2011). Friction: How radicalization happens to them and us. New York: Oxford University Press. McCauley, C., & Moskalenko, S. (2014). Toward a profile of lone wolf terrorists: What moves an individual from radical opinion to radical action. Terrorism & Political Violence, 26(1), 69-85. McGarrell, E. F., Chermak, S. M., & Freilich, J. D. (2007). Intelligence led policing as a framework for responding to terrorism. Journal of Contemporary Criminal Justice, 23(2), 142-158. McGarrell, E. F., Chermak, S. M., Weiss, A., & Wilson, J. M. (2001). Reducing firearms violence through directed police patrol. Criminology & Public Policy, 1(1), 119-148. McGloin, J. M. (2005a). Policy and intervention considerations of a network analysis of street gangs. Criminology & Public Policy, 4(3), 607-636. McGloin, J. M. (2005d). Street gangs and interventions: Innovative problem solving with network analysis. Washington, DC: Community Oriented Policing Services, Bureau of Justice Assistance, Office of Justice Programs, U.S. Department of Justice. McGloin, J. M., & Kirk, D. S. (2010a). An overview of social network analysis. Journal of Criminal Justice Education, 21(2), 169-181. McGloin, J. M., & Kirk, D. S. (2010b). Social network analysis. In A. R. Piquero & D. Weisburd (Eds.), Handbook of quantitative criminology (pp. 209-224). New York: Springer. McGloin, J. M., & Nguyen, H. (2012). It was my idea: Considering the instigation of cooffending. Criminology, 50(2), 463-494. McGloin, J. M., & Nguyen, H. (2014). The importance of studying co-offending networks for criminological theory and policy. In C. Morselli (Ed.), Crime and networks (pp. 13-27). New York: Routledge. McGloin, J. M., & O'Neill Shermer, L. (2009). Self-control and deviant peer network structure. Journal of Research in Crime and Delinquency, 46(1), 35-72. 231 McGloin, J. M., & Piquero, A. R. (2010). On the relationship between co-offending network redundancy and offending versatility. Journal of Research in Crime and Delinquency, 47(1), 63-90. McGloin, J. M., Sullivan, C. J., Piquero, A. R., & Bacon, S. (2008). Investigating the stability of co-offending and co-offenders among a sample of youthful offenders. Criminology, 46(1), 155-188. McIllwain, J. S. (1999). Organized crime: A social network approach. Crime, Law & Social Change, 32(4), 301-323. McPherson, M., Smith-Lovi, L., & Cook, J. M. (2001). Birds of a feather: Homophily in social networks. Annual Review of Sociology, 27, 415-444. Medina, R. M. (2014). Social network analysis: A case study of the Islamist terrorist network. Security Journal, 17(1), 97-121. Michael, G. (2012). Lone wolf terror and the rise of leaderless resistance. Nashville, TN: Vanderbilt University Press. Middleton, D. J., & Levi, M. (2005). The role of solicitors in facilitating ‘organized crime’: Situational crime opportunities and their regulation. Crime, Law & Social Change, 42(2/3), 123-161. Mitchell, R. G. (2002). Dancing at Armageddon: Survivalism and chaos in modern times. Chicago, IL: University of Chicago Press. Moody, J., McFarland, D., & Bender-deMoll, S. (2005). Dynamic network visualization. American Journal of Sociology, 110, 1206-1241. Morselli, C. (2001). Structuring Mr. Nice: Entrepreneurial opportunities and brokerage positioning in the cannabis trade. Crime, Law & Social Change, 35(3), 203-244. Morselli, C. (2003). Career opportunities and network-based privileges in the Cosa Nostra. Crime, Law & Social Change, 39(4), 383-418. Morselli, C. (2009). Inside criminal networks. New York: Springer. Morselli, C., & Giguere, C. (2006). Legitimate strengths in criminal networks. Crime, Law & Social Change, 45(3), 185-200. Morselli, C., Giguere, C., & Petit, K. (2007). The efficiency/security trade-off in criminal networks. Social Networks, 29(1), 143-153. Morselli, C., & Petit, K. (2007). Law-enforcement disruption of a drug importation network. Global Crime, 8(2), 109-130. 232 Morselli, C., & Roy, J. (2008). Brokerage qualifications in ringing operations. Criminology, 46(1), 71-98. Moskalenko, S., & McCauley, C. (2011). The psychology of lone-wolf terrorism. Counselling Psychology Quarterly, 24(2), 115-126. Mullins, S., & Dolnik, A. (2010). An exploratory, dynamic application of social network analysis for modelling the development of Islamist terror-cells in the West. Behavioral Sciences of Terrorism and Political Aggression, 2(1), 3-29. Natarajan, M. (2006). Understanding the structure of a large heroin distribution network: A quantitative analysis of qualitative data. Journal of Quantitative Criminology, 22(2), 171192. Nguyen, H., & McGloin, J. M. (2013). Does economic adversity breed criminal co-operation? Considering the motivation behind group crime. Criminology, 51(4), 833-870. Noble, D. F. (2004). Assessing the reliability of open source information. Paper presented at the 7th International Conference on Information Fusion, Stockholm, Sweden. http://www.fusion2004.foi.se/papers/IF04-1172.pdf Nonaka, I. (1994). A dynamic theory of organizational knowledge creation. Organization Science, 5(1), 14-37. O'Brien, S. P., & Haider-Markel, D. P. (1998). Fueling the fire: Social and political correlates of citizen militia activity. Social Science Quarterly, 79, 456-465. Papachristos, A. V. (2006). Social network analysis and gang research: Theory and methods. In J. F. Short & L. A. Hughes (Eds.), Studying youth gangs (pp. 99-116). Lanham, MD: Rowman & Littlefield Publishers. Papachristos, A. V. (2009). Murder by structure: Dominance relations and the social structure of gang homicide in Chicago. American Journal of Sociology, 115(1), 74-128. Papachristos, A. V. (2011). The coming of a networked criminology? In J. M. MacDonald (Ed.), Measuring crime and criminality: Advances in criminological theory (Vol. 17, pp. 101140). New Brunswick: Transaction Publishers. Papachristos, A. V., Meares, T. L., & Fagan, J. A. (2012). Why do criminals obey the law? The influence of legitimacy and social networks on active gun offenders. Journal of Criminal Law & Criminology, 102(2), 397-440. Parkin, W. S. (2012). Developing theoretical propositions of far-right ideological victimization. (PhD Dissertation), John Jay College, The Graduate Center, City University of New York, New York. Passas, N. (2003). Cross-border crime and the interface between legal and illegal actors. Security Journal, 16(1), 19-37. 233 Pedahzur, A., & Perliger, A. (2006). The changing nature of suicide attacks: A social network perspective. Social Forces, 84(4), 1987-2008. Perliger, A., & Pedahzur, A. (2011). Social network analysis in the study of terrorism and political violence. Political Science and Politics, 44(1), 45-50. Pettersson, T. (2003). Ethnicity and violent crime: The ethnic structure of networks of youths suspected of violent offenses in Stockholm. Journal of Scandinavian Studies in Criminology and Crime Prevention, 4(2), 143-161. Piazza, J. A. (2015). The determinants of domestic right-wing terrorism in the USA: Economic grievance, societal change and political resentment. Conflict Management and Peace Science. doi: 10.1177/0738894215570429 Pitcavage, M. (1998). Paper terrorism's forgotten victims: The use of bogus liens against private individuals and business. Washington DC: Anti-Defamation League. Retrieved from http://www.adl.org/mwd/privlien.asp Pitcavage, M. (1999). Old wine, new bottles: Paper terrorism, paper scams and paper "redemption". Washington DC: Anti-Defamation League. Retrieved from http://www.adl.org/mwd/redemption.asp Rapley, T. (2011). Some pragmatics of qualitative data analysis. In D. Silverman (Ed.), Qualitative Research (pp. 273-290). Thousand Oaks, CA: Sage. Ratcliffe, J. (2008). Intelligence-led policing. Portland: Willan Publishing. Reiss, A. J. (1986). Co-offender influences on criminal careers. In A. Blumstein, J. Cohen, J. A. Roth, & C. A. Visher (Eds.), Criminal Careers and Career Criminals (Vol. 2). Washington, DC: National Academy Press. Reiss, A. J., & Farrington, D. P. (1991). Advancing knowledge about co-offending: Results from a prospective longitudinal survey of London males. Journal of Criminal Law and Criminology, 82(2), 360-395. Reiss, A. J., & Tonry, M. (1993). Organizational crime. In A. J. Reiss & M. Tonry (Eds.), Beyond the Law: Crime in Complex Organizations (pp. 1-10). Chicago: University of Chicago Press. Robins, G. L., Elliott, P., & Pattison, P. E. (2001). Network models for social selection processes. Social Networks, 23(1), 1-30. Rogers, E. M. (2003). Diffusion of innovations (5th ed.). New York: Free Press. Ruggiero, V. (2003). Global markets and crime. In M. E. Beare (Ed.), Critical reflections on transnational organized crime, money laundering, and corruption (pp. 171-182). Toronto, CA: University of Toronto Press. 234 Sageman, M. (2004). Understanding terror networks. Philadelphia: University of Pennsylvania Press. Sageman, M. (2008). Leaderless jihad: Terror networks in the twenty-first century. Philadelphia University of Pennsylvania Press. Sampson, R. J., & Laub, J. H. (1995). Crime in the making: Pathways and turning points through life. Cambridge, MA: Harvard University Press. Sampson, R. J., & Laub, J. H. (2005). A life course view of the development of crime. The Annals of the American Academy, 602, 12-45. Sanchez, C. (2009). Sovereign citizens movement resurging: Resurgence of far-right movement reported. SPLC Intelligence Report, Issue 133. Montgomery, AL: Southern Poverty Law Center. Retrieved from http://www.splcenter.org/get-informed/intelligencereport/browse-all-issues/2009/spring/return-of-the-sovereigns Sarnecki, J. (2001). Delinquent networks: Youth co-offending in Stockholm. Cambridge: Cambridge University Press. Sarnecki, J., & Pettersson, T. (2001). Criminal networks in Stockholm. In M. W. Kleim, H. J. Kerner, C. Maxson, & E. G. M. Weitekamp (Eds.), The Eurogang paradox: Street gangs and youth groups in the U.S. and Europe (pp. 257-272). Amsterdam: Kluwer Academic. Shaw, C. R., & McKay, H. D. (1942). Juvenile delinquency and urban areas. Chicago: University of Chicago Press. Shelley, L. I. (2014). Dirty entanglements: Corruption, crime, and terrorism. Cambridge: Cambridge University Press. Skvoretz, J., & Faust, K. (1999). Logit models for affiliation networks. Sociological Methodology, 29(1), 253-280. Smith, B. L. (1994). Terrorism in America: Pipe bombs and pipe dreams. New York: State University of New York Press. Smith, B. L., Roberts, P., Gruenewald, J., & Klein, B. (2015). Patterns of lone actor terrorism in the United States. College Park, MD: National Consortium for the Study of Terrorism and Responses to Terrorism (START). Retrieved from https://www.start.umd.edu/pubs/START_ATS_PatternsofLoneActorTerrorismUS_Resea rchBrief.pdf?utm_source=START+Announce&utm_campaign=20fa35c0b0START_SpecialEditionCVE&utm_medium=email&utm_term=0_a60ca8c76920fa35c0b0-50035085 Snijders, T. A. B. (2005). Models for longitudinal network data. In P. J. Carrington, J. Scott, & S. Wasserman (Eds.), Models and methods in social network analysis (pp. 215-247). New York: Cambridge University Press. 235 Snijders, T. A. B. (2009). Longitudinal methods of network analysis. In R. A. Meyers (Ed.), Encyclopedia of complexity and system science (pp. 5998-6013). New York: Springer. Spaaij, R. (2010). The enigma of lone wolf terrorism: An assessment. Studies in Conflict & Terrorism, 33(9), 854-870. SPLC. (2001). Bob Schulz's We the People is the Don Quixote of Queensbury. SPLC Intelligence Report, Issue 104. Montgomery, AL: Southern Poverty Law Center. Retrieved from http://www.splcenter.org/get-informed/intelligence-report/browse-allissues/2001/winter/untaxing-america/don-quixote-of-queen SPLC. (2005). His 'straw man' free, a scammer finds the rest of him isn't: Patriots for profit. SPLC Intelligence Report, Issue 118. Montgomery, AL: Southern Poverty Law Center. Retrieved from http://www.splcenter.org/get-informed/intelligence-report/browse-allissues/2005/summer/patriots-for-profit SPLC. (2010). The sovereign: A dictionary of the peculiar. SPLC Intelligence Report, Issue 139. Montgomery, AL: Southern Poverty Law Center. Retrieved from http://www.splcenter.org/get-informed/intelligence-report/browse-allissues/2010/fall/sovereign-idioticon-a-dictionary-of-the Sullivan, B. A., & Chermak, S. M. (2012). The media's portrayal of product counterfeiting and financial crimes. International Journal of Comparative and Applied Criminal Justice, 36(4), 305-326. Sullivan, B. A., & Chermak, S. M. (2013). Product counterfeiting and the media: Examining news sources used in the construction of product counterfeiting as a social problem. International Journal of Comparative and Applied Criminal Justice, 37(4), 295-316. Sullivan, B. A., Chermak, S. M., Wilson, J. M., & Freilich, J. D. (2014). The nexus between terrorism and product counterfeiting in the United States. Global Crime, 15(3-4), 357378. Sullivan, B. A., Freilich, J. D., & Chermak, S. M. (2014). Financial Crime and Material Support Schemes Linked to al-Qa'ida and Affiliated Movements (AQAM) in the United States: 1990 to June 2014. College Park, MD: National Consortium for the Study of Terrorism and Responses to Terrorism (START). Retrieved from http://www.start.umd.edu/pubs/START_TEVUS_ECDB_Financial%20Crime%20and%2 0Material%20Support%20Schemes%20Perpetrated%20by%20Supporters%20of%20AQ AM_Nov2014.pdf Sullivan, B. A., Freilich, J. D., & Chermak, S. M. (forthcoming 2015a). Research Brief: Describing the American Far-Right Criminal Anti-Tax Movement. College Park, MD: National Consortium for the Study of Terrorism and Responses to Terrorism (START). 236 Sullivan, B. A., Freilich, J. D., & Chermak, S. M. (forthcoming 2015b). Research Highlight: Financial Crime Schemes Perpetrated by Far-Right Extremists in the United States: 19902013. College Park, MD: National Consortium for the Study of Terrorism and Responses to Terrorism (START). Sutherland, E. H. (1940). White-collar criminality. American Sociological Review, 5(1), 1-12. Sutherland, E. H. (1944). Is "white-collar crime" crime? American Sociological Review, 10(2), 132-139. Sutherland, E. H. (1949). White collar crime. New York: Dryden Press. Sutherland, E. H. (1983). White collar crime: The uncut version. Binghamton, NY: Vail-Ballou Press. Sutherland, E. H., Cressey, D. R., & Luckenbill, D. (1995). The theory of differential association. In N. J. Herman (Eds.), Deviance: A sympolic interactionist approach (pp. 64-68). Lanham, MD: General Hall (Rowman & Littlefield). Sykes, G. M., & Matza, D. (1957). Techniques of neutralization: A theory of delinquency. American Sociological Review, 22(6), 664-670. Tremblay, P. (1993). Searching for suitable co-offenders. In R. V. Clarke & M. Felson (Eds.), Routine activity and rational choice (pp. 17-37). New Brunswick, NJ: Transaction Publishers. U.S. Department of Treasury. (2014). Bogus Sight Drafts/Bills of Exchange Drawn on the Treasury. Washington, DC: U.S. Department of Treasury. Retrieved from http://www.treasurydirect.gov/instit/statreg/fraud/fraud_bogussightdraft.htm U.S. DOJ. (2012). HSBC Holdings Plc. and HSBC Bank USA N.A. Admit to Anti-Money Laundering and Sanctions Violations, Forfeit $1.256 Billion in Deferred Prosecution Agreement. Washington, DC: U.S. Department of Justice. Retrieved from http://www.justice.gov/opa/pr/hsbc-holdings-plc-and-hsbc-bank-usa-na-admit-antimoney-laundering-and-sanctions-violations U.S. GAO. (2008). Tax compliance: Businesses owe billions in federal payroll taxes. Washington, DC: U.S. Government Accountability Office. Retrieved from http://www.gao.gov/new.items/d08617.pdf van der Beken, T. (2004). Risky business: A risk-based methodology to measure organized crime. Crime, Law, and Social Change, 41(5), 471-516. van der Hulst, R. C. (2009). Introduction to social network analysis (SNA) as an investigative tool. Trends in Organized Crime, 12(2), 101-121. 237 van der Hulst, R. C. (2011). Terrorist networks: The threat of connectivity. In J. Scott & P. J. Carrington (Eds.), The SAGE handbook of social network analysis (pp. 256-270). Thousand Oaks, CA: Sage. van Duyne, P. C. (1993). Organized crime and business crime-enterprises in the Netherlands. Crime, Law, and Social Change, 19(2), 103–142. van Mastrigt, S. B., & Carrington, P. J. (2014). Sex and age homophily in co-offending networks: Opportunity or preference? In C. Morselli (Ed.), Crime and networks (pp. 2851). New York: Routledge. VERBI Software. (2015). MAXQDA: The art of text analysis: Reference manual. Sozialforschung: VERBI Software. Retrieved from http://www.maxqda.com/download/manuals/MAX11_manual_eng.pdf Waring, E. J. (2002). Co-offending as a network form of social organization. In E. Waring & D. Weisburd (Eds.), Crime and social organization: Advances in criminological theory (Vol. 10, pp. 15-29). New Brunswick, NJ: Transaction Publishers. Wasserman, S., & Faust, K. (1994). Social network analysis: Methods and applications. New York: Cambridge University Press. Weerman, F. M. (2011). Delinquent peers in context: A longitudinal network analysis of selection and influence effects. Criminology, 49(1), 253-286. Weisburd, D., Wheeler, S., Waring, E., & Bode, N. (1991). Crimes of the middle class: Whitecollar offenders in the federal courts. New Haven, CT: Yale University Press. Wellman, B. (1988). Structural analysis: From method and metaphor to theory and substance. In B. Wellman & S. D. Berkowitz (Eds.), Social structures: A network approach (pp. 1961). New York: Cambridge University Press. Wortley, R. (1997). Reconsidering the role of opportunity in situational crime prevention. In G. Newman, R. V. Clarke, & S. G. Shoham (Eds.), Rational choice and situational crime prevention (pp. 65-81). Aldershot: Ashgate. 238